Proton-transfer tautomerism of ?2-carbolines mediated by hydrogen-bonded complexes

Article abstract of DOI:10.1021/jp011031z

The carboxylic acid catalyzed excited-state amino-imino tautomerism for ?2-carboline (?2-CB) and its analogues has been investigated. Thermodynamics and microsolvation (i.e., stoichiometry of the complex formation) of various ?2-CB/acetic acid complexes in nonpolar solvents have been studied by means of absorption and emission titration experiments. Supplementary support of the stoichiometric ratio and structure for the hydrogen-bonding formation was provided by molecular design and syntheses of various ?2-CB analogues incorporating either only one hydrogen bonding site or dual hydrogen bonding sites where interplay between two sites are sterically prohibited. The results in combination with time-resolved measurements and theoretical approaches suggest the 1:2 ?2-CB/acetic acid complex with a structure of triple hydrogen bonding formation to be responsible for the excited-state proton-transfer tautomerism in cyclohexane. The proton transfer time is beyond the response limit of the detecting system of 15 ps, indicating that only a negligibly small geometry adjustment is required for the guest molecule (i.e., acetic acid) to a correct geometry, i.e., a hydrogen-bond relay configuration, for the triple proton transfer to proceed. In comparison, for the 1:1 ?2-CB/acetic acid non-hydrogen-bond relayed complexes, amino-imino tautomerism is prohibited during the excited-state lifetime, giving rise to a normal Stokes shifted emission. The results provide detailed ground-state thermodynamics of ?2-CB HB complexes as well as the dynamics of proton-transfer tautomerism mediated by the hydrogen-bonding structures.

Full text of DOI:10.1021/jp011031z

10674
J. Phys. Chem. B 2001, 105, 10674-10683
Proton-Transfer Tautomerism of â-Carbolines Mediated by Hydrogen-Bonded Complexes
Pi-Tai Chou,* Yun-I Liu, Guo-Ray Wu, Mei-Ying Shiao, and Wei-Shan Yu
Department of Chemistry, The National Chung-Cheng UniVersity, Chia Yi, Taiwan Republic of China
Chung-Chih Cheng and Chen-Pin Chang
Department of Chemistry, Fu Jen Catholic UniVersity, Shin Chuang, Taiwan
ReceiVed: March 19, 2001; In Final Form: July 5, 2001
The carboxylic acid catalyzed excited-state amino-imino tautomerism for â-carboline (â-CB) and its analogues
has been investigated. Thermodynamics and microsolvation (i.e., stoichiometry of the complex formation) of
various â-CB/acetic acid complexes in nonpolar solvents have been studied by means of absorption and
emission titration experiments. Supplementary support of the stoichiometric ratio and structure for the hydrogen-
bonding formation was provided by molecular design and syntheses of various â-CB analogues incorporating
either only one hydrogen bonding site or dual hydrogen bonding sites where interplay between two sites are
sterically prohibited. The results in combination with time-resolved measurements and theoretical approaches
suggest the 1:2 â-CB/acetic acid complex with a structure of triple hydrogen bonding formation to be responsible
for the excited-state proton-transfer tautomerism in cyclohexane. The proton transfer time is beyond the response
limit of the detecting system of 15 ps, indicating that only a negligibly small geometry adjustment is required
for the guest molecule (i.e., acetic acid) to a correct geometry, i.e., a hydrogen-bond relay configuration, for
the triple proton transfer to proceed. In comparison, for the 1:1 â-CB/acetic acid non-hydrogen-bond relayed
complexes, amino-imino tautomerism is prohibited during the excited-state lifetime, giving rise to a normal
Stokes shifted emission. The results provide detailed ground-state thermodynamics of â-CB HB complexes
as well as the dynamics of proton-transfer tautomerism mediated by the hydrogen-bonding structures.
1. Introduction
Spectroscopy and dynamics of host/guest hydrogen-bonding
(HB) type of excited-state proton-transfer reaction have long
receivedconsiderableattention.Prototypesinclude7-azaindoles,1-11
7-hydroxyquinolines,12-20 2-(2′-pyridyl)indoles,21-24 mono- and
di-pyrido2,3a carbazoles25-29 and â-carbolines30-40 where the
excited-state proton-transfer (ESPT) tautomerism is mediated
either by self-association or by adding guest molecules (includ-
ing solvents) forming host/guest types of HB complexes. From
the structural viewpoint, the proton donating and accepting sites
can be adjacent to each other so that a complex possessing intact
Figure 1. Conjugated dual hydrogen-bonding (CDHB) formation for
7AI, its associated redistribution of the electron density and the
corresponding ESDPT reaction in cyclohexane.
dual hydrogen bonds is formed through a perfect-fitted, 1:1
(host: guest) stoichiometric ratio in the ground state. For
example, the conjugated dual hydrogen bonding (CDHB)
effect41 in the case of 1:1 7-azaindole (7AI)/acetic acid complex
electron charge density in the pyrrolic ring was calculated to
mutually induces the π electron charge transfer from the pyrrolic
nitrogen to the pyridinic nitrogen (see Figure 1), resulting in
additional stabilization energy. As a result, an association
enthalpy of -14 kcal/mol has been reported.3 Upon electronic
excitation, the 7AI/guest CDHB complex undergoes an ultrafast
rate of double proton transfer (e.g., ∼1 ps-1 in the case of 7AI
dimer8,9), resulting in an imine-like tautomer fluorescence. In
contrast, for other type of systems the proton donating and
accepting sites may be far separated from each other. Thus, the
occurrence of ESPT possibly incorporates n (n g 2) guest
molecules. A prototype example is â-carboline (â-CB, see
Figure 2).42 From the molecular structure point of view â-CB
is composed of a π-electron deficient pyridinic ring fused to a
π-electron excessive indole ring and has been shown to reveal
interesting photophysical properties. In the ground state, the
be higher than its adjacent benzene and pyridine moieties. Upon
S0 f S1 (π, π*) excitation, the electronic charge density
migrates to the pyridinic nitrogen,30 resulting in a drastic change
in the acid-base property.31,32 In water and alcohols, equilibrium
among the neutral, cation and zwitterion was observed in the
excited state.32-34 In aprotic solvents such as dichloromethane
and dioxane, experiments have been performed to study the
photophysical properties of the â-CB/acetic acid complex.35-37
Unfortunately, the requirement of a large amount of acetic acid
due to the weaker HB association makes the thermodynamics
as well as microsolvation studies of the HB formation inacces-
sible. Furthermore, the requirement of a large fraction of acetic
acid creates an inhomogeneous local polar environment, which
induces an additional competitive channel based on the proto-
nation of â-CB. As a result, spectra and dynamics were
complicated by multiple fluorescence, consisting of neutral,
cation and zwitterionic species. Conversely, studies of â-CB/
* To whom correspondence should be addressed.
10.1021/jp011031z CCC: $20.00 © 2001 American Chemical Society
Published on Web 10/03/2001

Proton-Transfer Tautomerism of â-Carbolines
J. Phys. Chem. B, Vol. 105, No. 43, 2001 10675
column chromatography (eluent 1:1 (v/v) ethyl acetate: hex-
anes). H1NMR (CDCl3, 200 MHz), δ 4.36 (s, 3H), 6.91 (d, J )
6.8 Hz, 1H), 7.09 (d, J ) 6.8 Hz, 1H), 7.30-7.53 (m, 2H),
7.95 (d, J ) 6.8 Hz, 1H), 8.45 (d, J ) 6.8 Hz, 1H), 8.67 (s,
1H).
Synthesis of 2-Methyl-2H-â-carboline (C12H10N2, 2MCB).
2MCB was synthesized by the reaction of â-CB (0.1 g) and
methyl iodide (1.0 g) in THF under N2 atmosphere. NaOH (2.5
N,10 mL) was then added and the mixture was stirred for∼20
min to obtain product which was further purified by column
chromatography (eluent 1:4 (v/v) ethyl acetate: hexanes). H1-
NMR (CDCl3, 200 MHz), δ 4.37 (s, 3H), 7.24 (t, J ) 14.0 Hz,
1H), 7.62 (t, J ) 14.0 Hz, 1H), 7.64 (d, J ) 6 Hz, 1H), 7.89 (d,
J ) 8 Hz, 1H), 8.18 (d, J ) 6 Hz, 1H), 8.32 (d, J ) 6 Hz, 1H),
9.01 (s, 1H).
Synthesis of 1-Phenyl-1,2,3,4-tetrahydro-â-carboline (C17H16N2,
1PTCB). 1PTCB, a precursor to synthesize 1-phenyl-â-carboline,
was synthesized according to the Pictet-Spengler type of
reaction.46 Briefly, benzaldehyde (20 mmol) was added in one
portion to the stirred mixture containing tetra tryptamine (5
mmol in 50 mL dichloromethane) and trifluoroacetic acid (5
wt % in aqueous solution) at room temperature. The reaction
was allowed to proceed until tryptamine was completely
consumed as monitored by TLC (∼2-4 h). Upon completion
of the reaction, the reaction mixture was concentrated to dryness
under reduced pressure and then purified by column chroma-
Figure 2. Structures of the proton-transfer isomers and derivatives of
â-CB.
guest HB complexes in nonpolar, hydrocarbon solvents such
as cyclohexane are expected to provide more detailed insight
regarding the guest molecule coupled ESPT reaction. Recently,
ground-state HB formation and phototautomerism of 1-methyl-
â-carboline in cyclohexane titrated by the guest molecule have
been reported by Balo´n and co-workers.38-40 In their series of
papers, comprehensive studies have been performed to dif-
ferentiate the proton donor and acceptor guest molecules
affecting the HB formation as well as the corresponding excited-
state proton-transfer reaction.
In this study, acetic acid catalyzed excited-state amino-imino
tautomerism for â-CB and its analogues in nonpolar solvents
has been investigated. Thermodynamics and microsolvation (i.e.,
stoichiometry of the complex formation) of various â-CB/acetic
acid complexes in cyclohexane have been studied by means of
absorption and emission titration experiments. Owing to the
minimization of solvent perturbation the results have been
simulated by theoretical approaches. Supplementary support of
the stoichiometric ratio and structure for the hydrogen-bonding
formation was provided by molecular design and syntheses of
various â-CB analogues incorporating either only one hydrogen
bonding site or dual hydrogen bonding sites where interplay
between two sites are sterically prohibited. The results in
combination with dynamic studies render details in terms of
thermodynamics, microsolvation as well as ESPT dynamics of
â-CB/acetic acid HB complexes.
1
tography using ethyl acetate. H NMR (400 MHz, CDCl3), δ
2.91-3.27 (m, 4H), 5.58 (s, 1H), 7.13-7.23 (m, 4H), 7.29-
7.38 (m, 4H).
Synthesis of 1-Phenyl-â-carboline (C17H12N2, IPCB). A
solution of 1PTCB (2 mmol) in dioxane (75 mL) was refluxed
for 4 h in the presence of 2,3-dichloro-5, 6-dicyano-1, 4-ben-
zoquinone (4 mmol). The reaction mixture was cooled to room
temperature, and concentrated to dryness under reduced pressure
to obtain 1PCB. 1H NMR (400 MHz, CD3OD), δ 7.39 (m, H),
7.66-7.71 (m, 5H), 7.92-7.98 (m, 2H), 8.38 (d, J ) 8.16 Hz,
1H), 8.44 (d, J ) 5.92 Hz, 1H), 8.50 (d, J ) 5.92 Hz, 1H).
Synthesis of 1-Carboxylic-â-CB (C12H8N2O2, 1CCB): 1-Meth-
yl-â-carboline (0.5 mmol), potassium permanganate (2 mmol)
and dicyclohexyl-18-crown-6 (0.5 mmol) in benzene (20 mL)
were stirred in a ball mill for 72 h. The product potassium
benzoate was filtered along with manganese dioxide and then
was dissolved in 5% sodium hydride solution. After filtration
of manganese dioxide the aqueous solution was extracted with
ethyl ether to remove traces of crown ether followed by the
1
acidification to pH ) 6 to obtain 1CCB. H NMR (200 MHz,
CDCl3), δ 7.48-7.55 (m, 2H), 7.69 (d, J ) 6.4 Hz, 1H), 7.93
(t, J ) 7.6 Hz, 1H), 8.12 (d, J ) 2.2 Hz, 1H), 8.37 (d, J ) 2.2
Hz, 1H), 9.21 (s, 1H).
2. Experimental Section
Acetic acid (ACID, Merck Inc.) was purified according to
the previously described method.47 Cyclohexane (Merck Inc.)
was of spectrograde quality and was refluxed several hours over
calcium hydride under a nitrogen atmosphere and was trans-
ferred, prior to use, through distillation to the sample cell.
2.2 Measurements. Steady-state absorption and emission
spectra were recorded by a Cary 3E (Varian) spectrophotometer
and a F4500 (Hitachi) fluorimeter, respectively. Both wavelength-
dependent excitation and emission response of the fluorimeter
have been calibrated according to a previously reported method.47
Because the titration experiment is critical on the standpoint of
the interpreting hydrogen-bonding equilibrium (vide infra), we
have carefully performed our absorption and fluorescence
titration studies where each data point was taken by the average
of three to five measurements. The sensitivity for the absorption
2.1 Materials. Figure 2 depicts various structures of com-
mercially available and/or synthesized â-CB derivatives. â-CB
(Aldrich) and 1-methyl-â-carboline (1MCB,43 Aldrich) were
purified by column chromatography (eluent 1:3 (v/v) n-
hexane: ethyl acetate) followed by recrystallization from
acetonitrile. R-Carboline44 was synthesized according to the
previously reported method.45 The final product was twice
recrystallized from spectrograde ethanol and once from cyclo-
hexane. The purity of these compounds was checked by
fluorescence excitation spectra in dilute solution of <1.0 × 10-5
M where only monomer exists.
Synthesis of 9-Methyl-â-carboline (C12H10N2, 9MCB). 9MCB
was synthesized by adding sodium hydride (0.5 g) to the THF
solution containing â-CB (0.1 g) followed by the addition of
methyl iodide (0.15 g). The resulting product was purified by

10676 J. Phys. Chem. B, Vol. 105, No. 43, 2001
Chou et al.
Figure 4. Concentration-dependent absorption spectra of 9MCB (7.8
× 10-6 M) in cyclohexane at various ACID concentrations [C0g] of (a)
0, (b) 3.5 × 10-4, (c) 6.7 × 10-4, (d) 1.1 × 10-3, (e) 1.4 × 10-3, (f)
2.1 × 10-3, (g) 3.5 × 10-3, (h) 5.3 × 10-3, (i) 6.3 × 10-3, (j) 8.4 ×
10-3, and (k) 1.1 × 10-2 M. Insert: Plot of A0/(A0 - A) vs 1/[Cg] at
310 nm and its best linear least-squares fitted curve.
Figure 3. Concentration-dependent absorption spectra of â-CB (6.5
× 10-6 M) in cyclohexane at various ACID concentrations [C0g] of (a)
0, (b) 3.3 × 10-5, (c) 6.7 × 10-5, (d) 1.0 × 10-4, (e) 1.7 × 10-4, (f)
2.6 × 10-4, (g) 4.0 × 10-4, (h) 5.0 × 10-4, (i) 7.0 × 10-4, (j) 8.0 ×
10-4, and (k) 1.0 × 10-3 M. Insert A: Plot of 1/A350 vs 1/[Cg]2 assuming
the existence of 1:2 â-CB/ACID complexes exclusively in equilibrium,
B. Plot of [Cg]2/A350 vs [Cg]. Note that [Cg] depicted in the plot has
been corrected by using eq 4 from the originally prepared [C0g].
Similar notation of ACID was used in the rest of the figure captions.
in cyclohexane. Although self-dimerization is negligible the
initial concentration of â-CB, C0, was still prepared to be as
low as 6.5 × 10-6 M in this experiment to simplify further
derivation of equilibrium constants among various hydrogen-
bonded species. The formation of hydrogen-bonded complexes
between â-CB and ACID can be clearly shown by the growth
of an absorption shoulder at > 340 nm throughout the titration.
The appearance of isosbestic points located at ∼340 and 287
nm indicates the existence of equilibrium with common
intermediates.
measurement is approximately 2 × 10-4 in absorbance under
constant temperature (25 ( 0.1 °C) throughout the measurement.
Picosecond lifetime measurements were achieved by using either
the second or third harmonic of a femtosecond Ti-Sapphire
oscillator (82 MHz, 100 fs) as an excitation source. An
Edinburgh OB 900-L time-correlated single photon counter was
used as a detecting system. The fluorescence decays were
analyzed by the sum of exponential functions with an iterative
convolution method reported previously.48 This procedure allows
partial removal of the instrument time broadening and conse-
quently renders a temporal resolution of ∼15 ps.
According to the molecular structural analysis â-CB provides
two hydrogen-bonding sites. Therefore, the HB formation of
â-CB incorporating ACID molecules may stoichiometrically be
in the form of 1:1 and 1:2 â-CB/ACID complexes. The far
separation between two HB sites (see Scheme 1 for the
optimized structure) eliminates the possibility of forming a 1:1
cyclic dual HB â-CB/ACID complex that has been widely
accepted in 7AI analogues1-3,8-11 as well as in R-CB.52 We
first make an attempt to analyze the results incorporating
exclusively only one hydrogen-bonding site. Empirically, the
hydrogen-bonding strength between pyridinic nitrogen and
carboxylic hydrogen should be stronger than that between the
pyrrolic hydrogen and carbonyl oxygen due to their correspond-
ing higher acidity and basicity. The weaker hydrogen bond in
the pyrrolic hydrogen site can also be supported by the negligible
spectral change of carbazole in the same range of added ACID
concentrations (not shown here). Further support for the
correlation between hydrogen-bonding strength and acid-base
property of the proton donor and acceptor will be elaborated in
the later section. In a comparative study, upon increasing the
ACID concentration the model compound 9MCB where the
pyridinic nitrogen is the only hydrogen bonding site revealed a
small spectral change (<3 nm) in the S0 f S1 (ππ*, 330-360
nm) absorption region (see Figure 4). On the basis of the
structural similarity we therefore tentatively conclude that the
substantial change of the â-CB S0 f S1 (ππ*) absorption during
the ACID titration cannot be mainly attributed to the formation
of the 1:1 â-CB/ACID complex. The results on one hand may
indicate that there exists a single hydrogen-bond formation
between â-CB and ACID, but the redistribution of the electron
density induced by the single hydrogen bonding effect is small
so that the resulting S0 f S1 (ππ*) absorption remains nearly
unchanged during the titration. On the other hand, it is also
plausible that the formation constant of 1:1 â-CB/ACID complex
is small so that the concentration of the 1:1 HB complex is not
large enough to cause appreciable spectral change within the
2.3 Theoretical Calculations. Gaussian 98 Rev D.3 programs
were used to perform the ab initio calculation on the molecular
structure. Geometry optimizations for all structures were carried
out with the 6-31G(d, p) basis set at the Hartree-Fock (HF)
level. This basis set has proven to be suitable for the dimer (or
complex) formation incorporating hydrogen bonding forma-
tion.49 Hessians, and hence vibrational frequencies were also
performed to check whether the optimized geometrical structure
for those dimeric and complex forms is at an energy minimum,
transition state, or higher order saddle point. The directly
calculated zero-point vibrational energies were scaled by
0.918150 to account for the overestimation of vibrational
frequencies at the HF level. Because of the large HB systems
being studied, a counterpoise correction has been incorporated
to account for deviations from the basis-set superposition
(BSSE).51
3. Results
3.1 Absorption Titration Studies. Unlike R-carboline (R-
CB, see Figure 2), an isomer of â-CB with the interchange of
C(1) and N(2) positions, which exhibits significant dimerization
formation with a self-association constant of as high as ∼1.8
× 103 M-1 52-54
the absorption spectra (λmax ≈ 340 nm, ꢀ340 ≈
,
7690 M-1 cm-1) of â-CB reveal concentration independence
in cyclohexane, indicating negligible HB association. Molecular
geometry analyses presented in the following section of
theoretical approaches indicate that the formation of â-CB cyclic
dual hydrogen-bonded dimer is thermally unfavorable due to a
rise in the repulsion energy between two C(1)-hydrogen atoms
toward dimerization (vide infra). Figure 3 shows the absorption
spectra of â-CB upon adding various concentrations of ACID

Proton-Transfer Tautomerism of â-Carbolines
J. Phys. Chem. B, Vol. 105, No. 43, 2001 10677
SCHEME 1: Optimized Geometries Based on the HF/6-31G(d, p) Basis Set (in Å and Degrees, Only Critical Angles Are
Shown) for the Ground States of â-CB, 1CCB, 1PCB, and 1:2 â-CB/ACID THB Complex
added acetic acid concentrations. The former explanation can
be supported by a nonnegligible spectral change of 9MCB at
the S0 f Sn (n > 1) absorption region of ∼280-320 nm upon
the titration of acetic acid. On the basis of the formation of a
1:1 9MCB/ACID complex the relationship between the mea-
sured absorbance at a selected wavelength where both monomer
and 1:1 9MCB/ACID HB complex absorb (e.g., 310 nm) as a
function of the initially prepared ACID concentration, [Cg], can
be expressed by
â-CB/ACID complex, if it exists, should also have a similar Ka
value due to its structural similarity with respect to the 1:1
9MCB/ACID complex. The existence of a nonnegligible amount
of the 1:1 â-CB/ACID HB complex can also be supported by
a nonlinear plot of the reciprocal of absorbance at > 350 nm
vs 1/[Cg]2 (see insert A of Figure 3). Theoretically, the plot
should be linear if equilibrium exists exclusively between
uncomplexed â-CB and 1:2 â-CB/ACID HB complex.
The above results lead us to suggest a competitive equilibrium
between 1:1 and 1:2 â-CB/ACID complexes. In addition, one
has to consider the self-association of acetic acid due to its large
dimerization constant in cyclohexane. The corresponding mul-
tiple equilibrium in solution can thus be depicted as follows
A0
ꢀM
1
)
+ 1
(1)
(
)
[
]
A0 - A
ꢀM - ꢀC
Ka[Cg]
where ꢀM and ꢀC are molar extinction coefficients of the 9MCB
monomer and hydrogen-bonded complex monitored at a specific
wavelength, respectively. A detailed derivation of (1) has been
elaborated on ref 10b. The insert of Figure 4 shows the plot of
A0/(A0 - A) as a function of 1/[Cg] at a selected wavelength of
310 nm where the change of the spectral features is significant
upon adding ACID. Straight-line behavior supports the validity
of the assumption of 1:1 9MCB/ACID formation, and a linear
least-squares fit using eq. 1 deduces a Ka value of (2.2 ( 0.5)
× 102 M-1. It is therefore reasonable to predict that the 1:1
K1
â-CB + ACID y\z â-CB/ACID
K2
â-CB/ACID + ACID y\z â-CB/(ACID)2
2ACID yK3z (ACID)2
The acetic acid concentration used in this study is in an excess
amount relative to â-CB. Consequently, the equilibrium con-

10678 J. Phys. Chem. B, Vol. 105, No. 43, 2001
Chou et al.
centration of 1:1 â-CB/ACID complex ([C1:1]), 1:2 â-CB/ACID
complex ([C1:2]), ACID self-associated dimer and free ACID
([Cg]) simultaneously existing in the solution can be derived
and expressed by eq (2)-(4), respectively
[Cg]
K2
1
[C0]
1
)
[Cg]2 +
[C1:1] [C0]
[Cg] +
[Cg] +
(2)
(3)
[C0]K1
[Cg]2
1
1
1
)
[Cg]2 +
[C1:2] [C0]
K2[C0]
K1K2[C0]
0
0
(4K [C ] + 1) - 8K [C ] + 1
x
3
g
3
g
0
Figure 5. a. (s) Absorption and emission spectra of 1CCB in
cyclohexane, b. (- -) the excitation spectra of â-CB monitored at 500
nm by adding ACID (5.0 × 10-4 M) in cyclohexane.
[Cg] ) [Cg ] -
(4)
[
]
4K3
where [C0] denotes the initially prepared concentration of â-CB,
[C0] is the initially added acetic acid concentration, [Cg]
g
Scheme 1). Thus, the absorptivity of 1CCB at 350 nm can be
qualitatively used to simulate that of the 1:2 â-CB/ACID
complex, and a value of ꢀ350 was then estimated to be 3600
M-1 cm-1. Because C0 was prepared to be 6.5 × 10-6 M, a
value of 1945 ( 50 M-1 was deduced for K2, consistent with
the early assumption of 1/K2> [Cg] when [Cg] was prepared to
be within 2.4 × 10-5- 1.1 × 10-4 M in the titration study.
The result rationalizes the straight-line behavior obtained in
insert B of Figure 3. Our results also revealed a deviation from
straight-line behavior by plotting [Cg]2/A350 vs [Cg] when [Cg]
was prepared higher than 2 × 10-3 M, indicating that the
contribution from the [Cg]2 term becomes significant. However,
to prevent such a nonlinear plot and the possible cationic
formation due to the accumulation of local polar environ-
ments,35-37 low ACID concentrations of <1.0 × 10-3 M were
prepared throughout the study. According to the linear behavior
upon plotting [Cg]2/A350 vs [Cg], K1 was then extracted to be
180 ( 20 M-1 by dividing the slope with respect to the intercept,
which is on the same magnitude as that of the Ka value of ∼220
M-1 derived from the 1:1 9MCB/ACID complex.
3.2 Fluorescence Titration Study. Table 1 lists the steady-
state spectral properties of absorption and emission as well as
relaxation dynamics for â-CB, â-CB/ACID HB species and its
corresponding analogues in cyclohexane. Figure 6 shows the
fluorescence spectra as a function of the added ACID concen-
tration in cyclohexane. The uncomplexed â-CB (i.e., without
ACID added) shows a normal Stokes shifted emission maximum
at 360 nm of which the relaxation dynamics were well fit by
single-exponential kinetics with a lifetime of 2.8 ( 0.2 ns (Φf
≈ 0.2). Upon increasing the ACID concentration, excitation at
350 nm, where the â-CB/ACID HB complexes absorb pre-
dominantly, results in an increase of dual emission maxima at
365 and 510 nm, respectively. The 365 nm band is defined as
the F1′ component in order to distinguish it from the 360 nm
band (the F1 band) associated with the uncomplexed â-CB. The
F2 band (i.e., the 510 nm emission) exhibits emission-
wavelength independent relaxation dynamics, which can be well
fitted by a single-exponential decay rate of ∼1.7 × 108 s-1 (τf
≈ 5.9 ns), whereas the rise time is limited by the system
response of 15 ps. In a comparative study the synthesized imino-
tautomer analogue 2-methyl-2H-â-carboline (2MCB, see Figure
2) revealed a 538 nm emission (τf ≈ 2.41 ns) of which the
spectral features (i.e., the emission maximum, fwhm, etc., see
Figure 6) are similar to those of the F2 band. The results
unambiguously conclude that the 510 nm emission band is
attributed to the tautomer emission resulting from ACID
catalyzed N(9)-H to N(2) proton-transfer reaction. The ∼25
nm blue shift of the F2 band relative to 2MCB can be
represents the apparent concentration that has been corrected
by considering the self-association of acetic acid in cyclohexane.
In n-heptane the self-association constant of acetic acid K3 has
been reported to be 3.7 × 104 M-1 55 which was used in the
case of cyclohexane due to their similar solvent polarity. It has
been concluded that the growth of the absorbance at >340 nm
mainly originates from the 1:2 â-CB/ACID complex where the
resonance induction effect through a relay of hydrogen bonds
alters the electron density distribution. Incorporating the Beer-
Lambert law with an absorption cell of 1 cm path length, eq 3
associated with the existence of 1:2 â-CB/ACID complex can
be rewritten as
[Cg]2
1
1
1
)
[Cg]2 +
[Cg] +
(5)
A350
ꢀ350[C0]
ꢀ350K2[C0]
ꢀ350K1K2[C0]
In eq 5 A350 denotes the absorbance at 350 nm and is mainly
attributed to the absorbance of the 1:2 â-CB/ACID complex.
We have made an attempt to plot [Cg]2/A350 vs [Cg]. Interest-
ingly, instead of a quadratic plot between [Cg]2/A350 and [Cg]
which is expected from eq 5, the results reveal near straight-
line behavior (see insert B of Figure 3). A tentative but rational
explanation is that the coefficient 1/K2 in front of the [Cg] term
is much larger than [Cg] in the studied ACID concentrations so
that the quadratic term associated with [Cg]2 can be neglected.
[Cg] applied in the absorption titration experiment of â-CB is
typically within 2.4 × 10-5 and 1.1 × 10-4 M. To fulfill the
requirement of 1/K2> [Cg], K2 should be , 9000 M-1 to have
a negligible contribution of the [Cg]2 term. From another angle
of view one can rewrite eq 3 to {1/[C1:2] - 1/[C0]}[Cg]2 ) (1/
K2[C0])[Cg] + 1/K1K2[C0]. Throughout the titration study the
concentration of the 1:2 complex, [C1:2], may always be much
less than [C0] when both added ACID concentration and
association constants are relatively small. For this case 1/[C1:2
]
is expected to be . 1/[C0] throughout the titration experiment,
and the contribution of (1/[C0])[Cg]2 term in eq 5 can be
neglected. According to insert B of Figure 3 the best linear fit
gives a slope of 0.022 which is theoretically equivalent to
1/(ꢀ350K2[C0]) (see eq 5). Although ꢀ350 is not known at this
stage its value can be estimated qualitatively by the synthesis
of 1-carboxylic-â-carboline (1CCB, see Figure 2 for the
molecular structure). 1CCB exhibits an S0-S1 absorption band
(> 320 nm) similar to that of the excitation spectrum of â-CB/
ACID complex (see Figure 5, details will be discussed in the
section of the fluorescence titration). The results indicate that
the formation of two hydrogen bonds in 1CCB also mutually
induce each other through a carboxylic resonance structure (see

Proton-Transfer Tautomerism of â-Carbolines
J. Phys. Chem. B, Vol. 105, No. 43, 2001 10679
TABLE 1: Thermodynamic and Photophysical Properties of â-CB and Its Methylated Derivatives, ACID Complexes and â-CB
Analogues in Cyclohexane
absorption λmax (nm)
emission λmax (nm)
Φobs
0.20
0.27
0.017
0.30
τf (ns)
K1 or K2 (M-1)a (298 K)
â-CB
1MCB
2MCB
9MCB
1PCB
1CCB
1:2 â-CB/ACID
1:2 1MCB/ACID
1:1 9MCB/ACID
326, 340
325, 339
338, 346, 464
332, 346, 361
340, 351
338
344, 360
345, 359
538
365, 382, 402
367, 380
389
348, 365, 510
363, 495
367, 387, 407
2.80
2.58
2.41
2.24
1.33
2.05
0.034
0.14
347b
7.0 × 10-3,e
1.0 × 10-2,e
F1: 2.7, F1′: 3.8 F2: 5.9f
K1: 165c K2: 1.95 × 103
K1: 120c K2: 420
327, 342b
330, 345, 360
d
-
K1 : 220
a See text for the definition of K1 and K2. b Values were obtained from the fluorescence excitation spectrum. c Data were obtained from the
d
fluorescence titration experiment. K1 is equivalent to Ka defined in the text. e The quantum yield of imine tautomer emission. f See text for the
definition of F1, F1′, and F2.
Figure 7. (s) Fluorescence excitation spectrum of the â-CB monomer
(6.5 × 10-6 M) monitored at 370 nm. Fluorescence excitation spectra
of â-CB in 5 × 10-4 M acetic acid monitored at a. (‚‚‚‚‚) 390 nm, b.
(- - -) 530 nm.
Figure 6. (s) Fluorescence spectra of â-CB in cyclohexane at various
ACID concentrations [Cg0] of (a) 0, (b) 6.7 × 10-5, (c) 1.3 × 10-4, (d)
2.0 × 10-4, (e) 2.6 × 10-4, (f) 3.3 × 10-4, (g) 4.0 × 10-4, (h) 5.0 ×
10-4, (i) 6.0 × 10-4, (j) 8.0 × 10-4, and (k) 1.0 × 10-3 M. (- - -)
Absorption and emission spectra of 2MCB (1 × 10-5 M) in cyclohex-
ane. Insert: Plot of [Cg]2/F against [Cg] at the F2 band (520 nm).
where R is the instrument factor, including sensitivity, alignment,
etc., of the detecting system. As shown in Figure 6 the plot of
[Cg]2/F2 vs [Cg] revealed straight-line behavior. The results
further support the absorption titration study concluding that
the (1/ꢀ350[C0])[Cg]2term in eq 6 can be neglected due to its
small contribution in the studied ACID concentrations. Without
an absolute R value K2 cannot be extracted to make a
comparison with that obtained from the absorption titration
study. However, due to a better signal-to-noise ratio in the
fluorescence titration study a more reliable K1 value of ∼165
M-1 has been deduced by dividing the slope with respect to
the intercept.
The growth of the F1′ band upon increasing the acetic acid
concentration is intriguing. Figure 7a and b show the excitation
spectra of F1′ and F2 bands, respectively, upon adding 5.0 ×
10-4 M of acetic acid. In comparison, the excitation spectra
are different in which the excitation spectrum monitored at the
F1′ band is only slightly red shifted (<3 nm) with respect to
that obtained from uncomplexed â-CB. In contrast, the excitation
maximum of the F2 band is shifted by as large as ∼10 nm,
indicating that F1′ and F2 bands do not originate from a common
ground-state species. This viewpoint can be further supported
by a lifetime study. Detailed time-resolved studies indicate that
F1′ and F2 bands undergo different relaxation dynamics. In stead
of a well fitted single-exponential decay rate of 1.7 × 108 s-1
for the F2 band, an attempt to fit the F1′ band fitted by a single-
exponential decay component rendered a ø value that deviated
significantly from 1.0. Alternatively, it can be well fitted by a
double exponential decay, which is theoretically expressed as
rationalized by the hydrogen-bonding formation in the â-CB
(tautomer)/ACID complexes, which is commonly observed in
7AI and its corresponding analogues.3,10,11
For the 1:1 â-CB/ACID complex, it has been concluded in
an earlier section that the hydrogen-bonding pair consists of
carboxylic hydrogen (in ACID) and pyridinic nitrogen (in
â-CB). This configuration leads to the shortest distance between
the carbonyl oxygen and pyrrolic proton calculated to be as far
as ∼5.0 Å (vide infra). Thus, it is very unlikely that the double
proton transfer can take place through such a long-distance,
orientation-specified migration within the excited-state life span.
The system-response-limited (∼15 ps) rise of the F2 band also
discounts the possibility of the proton-transfer reaction through
the excited 1:1 â-CB/ACID complex in which an additional
carboxylic acid has to migrate in to form a proton relay
configuration, i.e., a precursor, for the proton transfer to proceed.
Detailed discussion will be elaborated in the following section.
Alternatively, because the increase of tautomer emission
intensity correlates well with an increase of the absorbance at
> 340 nm attributed to the 1:2 â-CB/ACID complex, it is more
plausible that the ground-state precursor responsible for the
proton transfer is ascribed to the 1:2 â-CB/ACID complex.
For a small concentration of the 1:2 â-CB/ACID complex
formation the observed tautomer fluorescence intensity is
theoretically proportional to the absorbance at e.g. 350 nm. Thus,
eq 5 can be rewritten to give eq 6
[Cg]2
1
1
1
F(t) ) A1e-k t + A2e-k t
) R
[Cg]2 +
[Cg] +
1
2
[
]
F2
ꢀ350[C0]
ꢀ350K2[C0]
ꢀ350K1K2[C0]
(6)
where A1 and A2 are the emission intensity at t ≈ 0 for the

10680 J. Phys. Chem. B, Vol. 105, No. 43, 2001
Chou et al.
4. Discussion
4.1 â-CB/Acetic Acid HB Structures. An intriguing ques-
tion is immediately raised regarding the structure of the 1:2
â-CB/ACID HB complex. Because of the highly unfavorable
entropy factor, large exothermicity is expected upon forming a
stable 1:2 â-CB/ACID complex. We here tentatively propose
that the structure of 1:2 â-CB/ACID complex be dominated by
a triple-hydrogen-bonded (THB) configuration in which two
ACID molecules forming hydrogen bonds with â-CB are linked
to each other by an additional hydrogen bond (see Scheme 1).
Although a direct experimental proof of the 1:2 â-CB/acetic
acid cyclic hydrogen-bonded complex in the solution phase is
not feasible at this stage, time-resolved studies and experiments
on â-CB derivatives have provided indirect evidence to support
this viewpoint. Global analyses of relaxation dynamics of the
tautomer fluorescence have been carried out over the studied
acetic acid concentrations. The results clearly revealed that
independent of acetic acid concentrations the relaxation dynam-
ics were well fitted by a single-exponential decay component.
No negative preexponential factor can be resolved, indicating
that the rate of acetic acid-catalyzed excited-state proton-transfer
Figure 8. Time-dependent fluorescence of â-CB in cyclohexane at
various ACID concentrations of a. 1.0 × 10-4 M, b. 5.0 × 10-3 M.
the fluorescence intensity was monitored at 410 nm with a spectral
bandwidth of ∼5 nm.
decay components 1 and 2, respectively. Although the ratio for
A1 vs A2 is concentration dependent, k1 and k2, within ACID
concentrations of 1.0 × 10-5-2.0 × 10-4 M, were found to be
constant, with two close decay components of (3.7 ( 0.1) ×
108 s-1(τ ≈ 2.7 ns) and (2.6 ( 0.2) × 108 s-1 (τ ≈ 3.8 ns),
respectively. The rise time for both components cannot be
resolved. The decay component with an τf of 2.7 ns is similar
to the decay dynamics of the uncomplexed â-CB. In addition,
the initial preexponential A1 factor decreases upon increasing
the ACID concentration. Thus, its assignment to the uncom-
plexed species is unambiguous. Accordingly, the 3.8 ns
component, of which the A2 factor increases as increasing the
ACID concentration, should be associated with the â-CB/ACID
complexes. Because the decay of 3.8 ns is much longer than
the system-response-limited rise time of the F2 band it is
impossible to assign the F1′ band to be the precursor for the F2
band. In addition, the assignment of the F1′ band to the â-CB
cationic emission can be discarded due to its emission maximum
of 365 nm that is drastically different from the reported cationic
emission maximum at 450 nm in other organic solvents
containing concentrated ACID.35-37 Accordingly, in the titration
study containing small ACID concentrations (<10-3 M) it is
reasonable to propose that the F1′ band results from an increase
of the 1:1 â-CB/ACID HB complex in which the proton transfer
is prohibited due to the far separation between ACID and
pyrrolic hydrogen.
In this study, when the ACID concentration was prepared to
be higher than 5 × 10-3 M, the contribution from an additional
decay component gradually appeared. Figure 8a shows the decay
of â-CB 410 nm emission (∆λ ≈ 5 nm) upon adding 1.0 ×
10-4 M ACID in cyclohexane, which reveals a major decay
component resulting from the 1:1 â-CB/ACID HB complex. In
comparison, upon adding the ACID concentration of as high
as 5 × 10-3 M a fast component with a decay rate of 7.1 × 109
s-1 (τf ≈ 140 ps, see Figure 8b) appeared. The amplitude of
the 140 ps decay component increases as the ACID concentra-
tion increases. In the strictly nonpolar, hydrocarbon solvents
such as cyclohexane the assignment of this fast decay component
to the protonated pyridinic nitrogen (â-CB)/carboxylate anion
(ACID) ionpair emission is reasonable, which can be distin-
guished from the solvated â-CB cationic emission in the highly
concentrated ACID and polar environment.35-37 In the higher
concentration of > 5% ACID by weight, similar to â-CB titrated
by ACID in other organic solvents, a solvated cationic emission
maximum at 450 nm gradually appeared. The high ACID
concentration apparently makes the interpretation of thermo-
dynamics (i.e., equilibrium) and dynamics of â-CB HB com-
plexes more complicated and was hence avoided in this study.
reaction is beyond our current system response of ∼15 ps-1
.
For the 1:2 â-CB/ACID noncyclic complex where two acid
molecules are individually hydrogen bonded to the pyridinic
nitrogen and pyrrolic hydrogen, respectively with a lack of relay
configuration ESPT requires a diffusive migration of the acid
molecules, forming a precursor for the proton transfer to take
place. The time scale of diffusive migration, depending on the
solvent viscosity, normally takes place within several hundred
picoseconds or longer in the low and medium viscous solvents
(for example, see the cases of 7-azaindole in alcohols5,6). Similar
standpoint holds for the 1:2 â-CB/ACID noncyclic complex
where the acetic acid molecule is hydrogen bonded with
pyridinic nitrogen and the other acetic acid simultaneously. On
the other hand, the possibility of proton-transfer mechanism
where a 1:2 noncyclic â-CB/ACID complex is catalyzed by
another guest (i.e., acetic acid) molecule is also considered. In
this case the rate of proton transfer can be expressed as kpt
)
kd[guest] where the upper limit of kd is assumed to be the
diffusion-controlled rate. This proposal is discounted in the case
of â-CB titrated by acetic acid. For a simplified approach, the
tautomer emission was observed at the ACID concentration as
low as 5 × 10-5 M. Taking kd to be ∼2 × 1010 M-1 s-1 in
cyclohexane kpt is calculated to be ∼106 s-1, which is too slow
to compete with other nonproton-transfer decay processes and
is contradictory to the observed ACID concentration indepen-
dent, system-response-limited rise time of the tautomer emission.
It is thus conceivable for us to propose the formation of a 1:2
â-CB/ACID cyclic complex possessing a THB structure where
ESPT takes place through a conduit of relayed hydrogen bonds.
A similar 1:2 (host/guest) cyclic THB complex has been
proposed in the case of 7-hydoxyquinoline titrated by metha-
nol.16 For the 1:2 â-CB/ACID THB complex it is further
proposed that the proton donor and acceptor sites are mutually
induced through a conjugated triple hydrogen-bonding (CTHB)
effect, resulting in a redistribution of electron density from
indole to pyridine ring. As a result, a significant change in the
spectral properties is expected on the S0 f S1 (ππ*) absorption,
consistent with the experimental results. This viewpoint is
similar to that proposed in the case of 1:1 7AI/guest cyclic HB
complexes in which the CDHB effect results in a drastic change
of the absorption spectral features.1-3 The linkage of the third
hydrogen bond in combination with the CTHB effect stabilizes

Proton-Transfer Tautomerism of â-Carbolines
J. Phys. Chem. B, Vol. 105, No. 43, 2001 10681
in the case of R-CB and 7AI (see Figure 2 for the comparison).
Certainly, as proposed in several literature,35-40 there is other
possibility of forming a zwitterionic tautomer form that might
possess a large dipolar change and thus can be stabilized in the
polar solvents. Ab initio calculation focusing on the zwitterion
formation incorporating the solvation effect is complicated and
was not performed in this study. The calculation also reveals a
far separated distance (4.157 Å, see Scheme 1) between proton
donor (the indolic hydrogen) and acceptor (the pyridinic
nitrogen) sites in â-CB (amino form). All attempts to resolve
the dimeric form gave rise to imaginary vibrational frequencies,
indicating that no stable dimeric form exists based on the ab
inito approach (HF/6-31G(d, p) method). The failure in obtaining
a geometry optimized â-CB CDHB dimeric form is possibly
due to the buildup of an enormously large repulsion energy
between two C(1)-H hydrogens when two â-CB molecules
approach each other in a C2V symmetry, rationalizing the failure
in detecting any â-CB self-association in both absorption and
fluorescence titration studies.
Figure 9. Fluorescence spectra of 1MCB in cyclohexane titrated by
ACID in which the ACID concentration is prepared to be (a) 0, (b) 1.1
× 10-4, (c) 2.1 × 10-4, (d) 3.5 × 10-4, (e) 4.2 × 10-4, (f) 5.3 × 10-4
,
(g) 6.5 × 10-4 and (h) 8.0 × 10-4 M Insert: A concentration of as
high as 1.0 × 10-2 M has to be added in order to obtain an optimum
tautomer emission intensity.
Unlike 7AI/ACID and R-CB/ACID hydrogen-bonded com-
plexes incorporating a strong conjugated dual hydrogen-bonding
effect, the formation of a 1:1 â-CB/ACID hydrogen-bonded
complex, if it exists, should involve only a single hydrogen bond
even though acetic acid possesses bifunctional hydrogen bonding
sites. On the other hand, the far separation between proton donor
and acceptor sites in â-CB leads to a favorable enthalpy factor
toward the formation of 1:2 â-CB/ACID HB complexes. For
the 1:2 â-CB/ACID complex possessing a THB configuration
(see Scheme 1) the enthalpy of association ∆Hac was calculated
to be -20.78 kcal/mol. The strong triple hydrogen bond and
its corresponding CTHB effect should be responsible for the
highly exothermic reaction toward forming a relay type of HB
complex. Although such a configuration requires a specific
configuration of hydrogen bonds, the large exothermicity
compensates a less favorable entropy effect. As a result, the
free energy was calculated to be 2.5 kcal/mol lower than that
of the free â-CB and ACIDs. Conversely, the 1:2 â-CB/ACID
nonproton-relayed DHB complex is calculated to be ∼9 kcal/
mol less stable in enthalpy than that of the THB complex, and
hence a free energy of 4.7 kcal/mol higher than that of the THB
complex. The calculation also indicates that the 1:1 â-CB
(pyridinic nitrogen)/ACID (carboxylic hydrogen) HB complex
is more stable than the 1:1 â-CB (pyrrolic hydrogen)/ACID
(carbonyl oxygen) HB complex by 3.5 kcal/mol. The result
predicts the hydrogen-bonding site in the 1:1 ACID/â-CB
complex to be located at the pyridinic nitrogen, consistent with
the experimental result. Further evidence can be given by the
similar enthalpy of association calculated between 1:1 â-CB
(pyridinic nitrogen)/ ACID (carboxylic hydrogen) (∆Hac ≈ -7.2
kcal/mol) and 1:1 9MCB/ACID (∆Hac ≈ -7.0 kcal/mol) HB
complexes. Note that 9MCB provides only one hydrogen-
bonding site at pyridinic nitrogen. This site selective hydrogen-
bonding formation can be rationalized qualitatively through the
acid-base property in the formation of a hydrogen bond.
According to the pH titration experiment the carboxylic
hydrogen (pKa ≈ 4.74) is much more acidic than that of the
pyrrolic hydrogen (pKa ≈ 14.556) in â-CB. Conversely, the
pyridinic nitrogen (pKb ≈ 8.958) in â-CB is more basic than
the carbonyl oxygen (CdO, pKb ≈ 22.258). Treating the
hydrogen-bonding formation as an acid (proton donor)-base
(proton acceptor) type of reaction, the sum of pKa and pKb is
equivalent to -logKeq where Keq denotes an equilibrium constant
for the acid-base reaction. Hence, under negligible influence
of the steric effect a lower pKa + pKb value indicates a large
the 1:2 â-CB/ACID complex, explaining the relatively large
K2 value derived from the experimental results.
A strong support for the proposed CTHB effect can also be
given by the synthesis of 1PCB (see Figure 2). The geometry
optimized (6-31G(d, p)) structure of 1PCB reveals the C(1)
substituted phenyl ring to be ∼42.5° (see Scheme 1) with respect
to the molecular plane of parent â-CB due to its steric effect
with the R hydrogen. Although 1PCB provides dual hydrogen
bonding sites the bulky phenyl substitution blocks any possible
linkage between two ACID molecules (i.e., the formation of a
third hydrogen bond). Experimentally, upon increasing the
ACID concentration the results show negligible spectral change
on the S0 f S1 (ππ*, > 320 nm) absorption except for the
increase of the absorptivity on the S0 f Sn (ππ*, n >1)
absorption at 280-320 nm (not shown here), an indication of
the possible 1:1 1PCB/ACID HB complex formation. Upon
excitation the tautomer emission was completely obscured
throughout the ACID titration experiment. Further evidence can
be given by 1MCB where the methyl group, to a certain extent,
hinders the formation of 1:2 1MCB/ACID THB complex. To
achieve a similar spectral change as that of the 1:2 â-CB/ACID
THB complex (Figure 3) more ACID concentration (i.e., 5.0 ×
10-3 M) has to be added in the ACID titration study for 1MCB.
Accordingly, a much smaller K2 value (∼420 M-1) for the 1:2
1MCB/ACID formation has been deduced from the absorption
titration study. The fluorescence titration study of 1MCB shown
in Figure 9 revealed that the F2 band was obscured at similar
added ACID concentrations as in Figure 6. Higher ACID
concentrations have to be added to obtain optimum F2 emission
(see insert of Figure 9). The result clearly indicates that a steric
hindrance introduced by the phenyl or methyl group at the C(1)
position either prohibits or hampers the THB formation. Further
supplementary supports of the proposed HB structures are
provided by the theoretical approaches discussed in the follow-
ing section.
4.2 Theoretical Approach. The geometrically optimized (6-
31G(d,p)) â-CB exists predominantly as an amino-like form
(see Figure 2 and Scheme 1) which is more stable than its
corresponding imino tautomer by ∼24 kcal/mol. In comparison,
the difference in free energy between amino and imino forms
is calculated to be ∼11.8 and 12.3 kcal/mol for R-CB and 7AI,
respectively. The higher endergonic â-CB (amino form) f â-CB
(imino form) tautomerism can be rationalized by destroying two
aromatic rings upon forming the tautomer (i.e., imino) form,
whereas it only requires the destruction of single aromaticity

10682 J. Phys. Chem. B, Vol. 105, No. 43, 2001
Chou et al.
acid-base equilibrium constant, and the reaction favors the
product formation, consistent with both experimental and
theoretical approaches for the hydrogen-bonding structure of
the â-CB/ACID complex. Note that this conclusion is qualitative
because the hydrogen-bonded complex is in its local minimum
energy where the proton is neither accepted nor donated
completely. In addition, the correlation was made where the
difference in pKa + pKb between various hydrogen-bonding
configurations is generally quite large (normally . 5). Once
the difference becomes small the correlation may be discounted
due to an uncertainty by applying the pKa (or pKb) value
obtained in the aqueous solution to a nonpolar (or gas)
environment. When the proton affinity between donor and
acceptor is close enough, the formation of an unusually strong
hydrogen bond has been reported in several enzymatic reactions
and subsequently verified by theoretical approaches,60-62 for
which the aforementioned empirical correlation may no longer
be valid. Fortunately, the proton affinity for various functional
groups applied in the case of â-CB hydrogen-bonded complex
is substantially different from each other, excluding the pos-
sibility of an unusually strong hydrogen-bonding formation.
SCHEME 2: Proposed Triple-proton-transfer
Mechanism for the 1:2 â-CB/ACID THB Complex.
*Indicates the Excited Singlet State (1ππ*). Critical Bond
Distances and Angles Are in Å and Degrees
1CCB. Accordingly, the rate of proton transfer may be too slow
to compete with other relaxation dynamics in the excited state.
In comparison, the proton-relay configuration in the 1:2 â-CB/
ACID THB complex is relatively much more flexible than
1CCB. As a result an ESPT mechanism incorporating triple
proton transfer is proposed and depicted in Scheme 2. The
π-electron excitation of the 2:1 â-CB (-N-)/ACID (-OH)
THB complex mutually induces a charge redistribution of â-CB,
resulting in an increase of the acidity and basicity for the pyrrolic
hydrogen and pyridinic nitrogen, respectively. A ∆pKa (pKa*-
pKa) as much as -5.5 and 6.5 for pyrrolic hydrogen and
protonated pyridinic nitrogen, respectively, has been reported.32
The COOH‚‚‚‚‚N(2) hydrogen bond of ∼1.87 Å is much shorter
than that of 2.14 Å for the CdO‚‚‚‚‚H-N(9) site (see Scheme
1). Whereas for the 1:2 â-CB (tautomer)/ACID THB complex
the relative HB distance is reversed with COO-H‚‚‚‚‚(9) of
1.85 Å < CdO- - -H-N(2) of 2.15 Å (see Scheme 2). Thus, a
triggering step for the ESDPT may incorporate the proton
transfer through a preexisting â-CB(-N(2)-)/ACID(-OH)
hydrogen bond, which leads to an increase of the basicity, i.e.,
a buildup of partial negative charge density, at the carbonyl
oxygen site of the other ACID molecule through a proton relay.
This in combination with the partial positive charge of the
pyrrolic proton creates an electrostatic attraction, which acts as
a driving force for a slight adjustment (or displacement) of both
â-CB and ACID to a correct conformation. Thus, the second-
step in the proton transfer can take place through a lowest
potential energy surface. For this case, two ACID molecules
linked with a hydrogen bond act as a proton bridge to achieve
an autocatalytic ESPT process. This proposed mechanism
incorporates hydrogen-bonding-relay coupled proton transfer (or
tunneling) dynamics and, hence, is expected to reveal deuterium-
isotope dependence. Unfortunately, upon deuteration on both
â-CB and ACID, the rise time of the tautomer emission was
still beyond the response of our photon counting system of ∼15
ps. Further study focusing on the â-CB proton-transfer dynamics
in a pico-femtosecond time scale should be performed to verify
the proposed mechanism.
4.3 Proton-Transfer Dynamics. Both experimental and
theoretical approaches have drawn the conclusion that the
increase of the proton-transfer tautomer emission is mainly
associated with the formation of a 1:2 â-CB/ACID THB
complex. On the basis of a nanopicosecond time-resolved study,
the rate of excited-state proton transfer for the 1:2 â-CB/ACID
THB complex was estimated to be > 15 ps-1 at room
temperature. In the cases of the 7AI dimer or 7AI/ACID HB
complex where cyclic dual hydrogen bonds are intrinsically
formed, the double proton transfer in the excited state may only
require a negligibly small displacement of the hydrogen atom
and/or molecular skeleton. The rate of such a cooperative proton-
transfer reaction, either taking place stepwise or simultaneously,
is fast (∼1 ps-1) and perhaps mainly dominated by the proton
tunneling mechanism.8,9 In comparison, the occurrence of
ESDPT in the 1:2 â-CB/ACID THB complex is intriguing from
the viewpoint of the proton-transfer dynamics. Could the THB
complex be in a perfect geometrical configuration so that triple
proton transfer occurs instantaneously without any diffusive
adjustment? Although this question still remains unanswered
at this stage it is believed that a small geometrical adjustment
(or even solvent perturbation) associated with the hydrogen-
bonding-relay configuration might be crucial in dealing with
the triple-hydrogen-bond relay type of proton transfer. This
viewpoint has been supported by the synthesis of 1CCB. As
shown in Scheme 1, the geometry optimized structure of 1CCB
based on 6-31G(d, p) basis sets reveals the formation of
intramolecular dual hydrogen bonds in a relay type of config-
uration. Such a configuration has been spectrally identified by
its S0 f S1 (ππ*) absorption spectrum which correlates well
with the fluorescence excitation spectra of the imine-tautomer
emission resulting from the THB complex (see Figure 5).
Surprisingly, despite the relay type of dual hydrogen bonds
where an excited state proton-transfer tautomerism is expected,
only a normal Stokes shifted emission maximum at 390 nm
was observed in 1CCB (see Figure 5). Detailed analyses indicate
that the CdO‚‚‚‚‚H-N(9) hydrogen bond of ∼2.3 Å is much
longer than that of 2.1 Å for the O-H‚‚‚‚‚N(2) site (see Scheme
1). Thus, ESPT may require a slight adjustment of the
CdO‚‚‚‚‚H-N hydrogen bond to an optimum distance for the
proton transfer reaction to proceed. Such an adjustment from
an “incorrect” to a “correct” geometry may incorporate a large
energy barrier due to the rigidity of the carboxylic group in
5. Conclusion
In conclusion, we have studied thermodynamics of â-CB/
ACID hydrogen-bonded complexes by means of absorption,
fluorescence titration experiments. In cyclohexane, the ground-

Proton-Transfer Tautomerism of â-Carbolines
J. Phys. Chem. B, Vol. 105, No. 43, 2001 10683
(20) Chou, P. T.; Wei, C. Y.; Wang, C. R. C.; Hung, F. T.; Chang, C.
P. J. Phys. Chem. A 1999, 103, 1939.
(21) Rodr´ıguez-Prieto, F.; Mosquera, M.; Novo, M. J. Phys. Chem. 1990,
94, 8536.
(22) Herbich, J.; Hung, C. Y.; Thummel, R. P.; Waluk, J. J. Am. Chem.
Soc. 1996, 118, 3508.
(23) Kyrychenko, A.; Herbich, J.; Wu, F.; Thummel, R. P.; Waluk, J.
J. Am. Chem. Soc. 2000, 122, 2818.
(24) Rios Rodr´ıguez, M. C.; Mosquera, M.; Rodr´ıguez-Prieto, F. J. Phys.
Chem. A 2001, ASPA.
(25) Kyrychenko, A.; Herbich, J.; Izydorzak, M.; Wu, F.; Thummel, R.
P.; Waluk, J. J. Am. Chem. Soc. 1999, 121, 11 179.
(26) Marks, D.; Zhang, H.; Borowicz, P.; Waluk, J.; Glasbeek, M. J.
Phys. Chem. A 2000, 104, 7167.
(27) De Valle, J. C.; Dominguez, E.; Kasha, M. J. Phys. Chem. A 1999,
103, 2467.
(28) Kyrychenko, A.; Stepanenko, Y.; Waluk, J. J. Phys. Chem. A 2000,
104, 9542.
(29) Waluk, J. Conformational Aspects of Intra- and Intermolecular
Excited State Proton Transfer In. Conformational Analysis of Molecules in
Excited States; Waluk, J., Ed.; Wiley-VCH, 2000, and references therein.
(30) Draxler, S.; Lippitsch, M. E. J. Phys. Chem. 1993, 97, 11 493-
11 496.
(31) Ghiggino, K. P.; Skilton, P. F.; Thistlethwaite, P. J. J. Photochem.
1985, 31, 113.
(32) Balo´n, M.; Hidalgo, J.; Guardado, P.; Mun˜oz, M. A.; Carmona, C.
J. Chem. Soc. Perkin Trans. 2 1993, 99.
(33) Varela, A. P.; Miguel, M. da G.; Macanita, A. L.; Burrows, H. D.;
Becker, R. S. J. Phys. Chem. 1995, 99, 16 093.
state equilibrium for â-CB titrated by trace amounts of ACID
consists of uncomplexed â-CB, 1:1 â-CB/ACID single HB and
1:2 â-CB/ACID complexes. Specific hydrogen-bonding sites
and structures of the HB complex have been deduced by
applying various derivatives of â-CB incorporating either only
a hydrogen site or dual hydrogen bonding sites where interplay
between two sites are sterically prohibited. The results in
combination with time-resolved measurements and theoretical
approaches suggest that the 1:2 â-CB/ACID complex consists
of a triple-hydrogen-bonding formation where fast ESPT (<15
ps-1) takes place through a conduit of relayed hydrogen bonds.
For the 1:1 â-CB/ACID non-hydrogen-bond relayed complexes,
amino-imino tautomerism is prohibited during the excited-state
lifetime, giving rise to a normal Stokes shifted emission. The
results provide detailed ground-state thermodynamics and
dynamics of excited-state proton transfer tautomerism for â-CB/
ACID HB complexes. Focus on the molecular dynamics
approach of the â-CB/ACID hydrogen-bonding formation in
cyclohexane is currently in progress in attempt to render more
substantial evidence for the structures of â-CB/ACID HB
complexes in cyclohexane.
Acknowledgment. Start-up support from the National
Chung-Cheng University is graciously acknowledged. This work
was supported by the National Science Council, Taiwan, R.O.C.
(Grant No. NSC 89 -2113-M-194-009). We thank the National
Center for High-Performance Computing, Taiwan for the use
of their facility.
(34) Dias, A.; Varela, A. P.; Miguel, M. da G.; Macanita, A. L.; Becker,
R. S. J. Phys. Chem. 1992, 96, 10 290-10 296.
(35) Reyman, D.; Pardo, A.; Poyato, J. M. L. J. Phys. Chem. 1994, 98,
10 408-10 411.
(36) Reyman, D.; Vin˜as, M. H.; Poyato, J. M. L.; Pardo, A. J. Phys.
Chem. A 1997, 101, 768.
(37) Reyman, D.; Vin˜as, M. H. Chem. Phys. Lett. 1999, 301, 551.
(38) Balo´n, M.; Mun˜oz, M. A.; Guardado, P.; Carmona, C. Photochem.
Photobiol. 1996, 64, 531.
(39) Balo´n, M.; Carmona, C.; Guardado, P.; Mun˜oz, M. A. Photochem.
Photobiol. 1998, 67, 414.
(40) Carmona, C.; Gala´n, M.; Angulo, G.; Mun˜oz, M. A.; Guardado,
P.; Balo´n, M. Phys. Chem. Chem. Phys. 2000, 2, 5076.
(41) Chou, P. T.; Wei, C. Y.; Hung, F. T. J. Phys. Chem. B 1997, 101,
9119.
(42) Norharman is an alternative name for â-CB, which is commonly
used in the literature. However, the name of â-CB is preferred due to a
series of synthesized â-CB derivatives presented in this study. An IUPAC
name 9H-pyrido[3,4-b]indole has also been used in Chemical Abstracts.
(43) Alternatively, harmane is a common name for 1-Methyl-â-carboline.
(44) 1-Azacarbazole has been commonly used in the literature. In this
study the alternative name R-CB was applied in order to have a clear
comparison with â-CB.
(45) Stephenson, L.; Warburton, W. K. J. Chem. Soc. 1970, 1355.
(46) Cox, E. D.; Cook, J. M. Chem. ReV. 1995, 95, 1797.
(47) Wei, C. Y.; Yu, W. S.; Chou, P. T.; Hung, F. T.; Chang, C. P.;
Lin, T. C. J. Phys. Chem. B 1998, 102, 1053.
(48) Demas, J. N.; Crosby, G. A. J. Phys. Chem. 1971, 75, 991.
(49) (a) Wang, J.; Boyd, R. J. Chem. Phys. Lett. 1996, 259, 647. (b)
Wang, J.; Boyd, R. J. J. Phys. Chem. 1996, 100, 16 141.
(50) Wong, M. W.; Wiberg, K. B.; Frisch, M. J. J. Am. Chem. Soc.
1992, 114, 1645.
(51) (a) Stewart, J. J. P. J. Comput. Chem. 1989, 10, 221. (b) Buemi,
G.; Zuccarello, F.; Raudino, A. THEOCHEM 1988, 41, 379.
(52) Chang, C. P.; Yen, F. H.; Chou, P. T.; Wei, C. Y. J. Chin. Chem.
Soc. 1996, 43, 463.
(53) Chang, C. P.; Shabestary, N.; El-Bayoumi, M. A. Chem. Phys. Lett.
1980, 75, 107.
References and Notes
(1) Taylor, C. A.; El-Bayoumi, M. A.; Kasha, M. Proc. Natl. Acad.
Sci. U.S.A. 1969, 63, 253.
(2) Ingham, K. C.; El-Bayoumi, M. A. J. Am. Chem. Soc. 1974, 96,
1674.
(3) (a) Chang, C. P.; Hwang, W. C.; Kuo, M. S.; Chou, P. T.; Clements,
J. H. J. Phys. Chem. 1994, 98, 8801. (b) Chou, P. T.; Wei, C. Y.; Chang,
C. P.; Chiu, C. H. J. Am. Chem. Soc. 1995, 117, 7259. (c) Chou, P. T.;
Wei, C. Y.; Chang, C. P.; Kuo, M. S. J. Phys. Chem. 1995, 99, 11 994.
(4) Koijnenberg, J.; Huizer, A. H.; Varma, C. A. O. J. Chem. Soc.
Faraday Trans. 2 1988, 84 (8), 1163.
(5) Moog, R. S.; Maroncelli, M. J. Phys. Chem. 1991, 95, 10 359. (b)
Chapman, C. F.; Maroncelli, M. J. Phys. Chem. 1992, 96, 8430. (c) Mente,
S.; Maroncelli, M. J. Phys. Chem. A 1998, 102, 3860.
(6) (a) Negrerie, M.; Gai, F.; Bellefeuille, S. M.; Petrich, J. W. J. Phys.
Chem. 1991, 95, 8663. (b) Gai, F.; Chen, Y.; Petrich, J. W. J. Am. Chem.
Soc. 1992, 114, 8343. (c) Chen, Y.; Rich, R. L.; Gai, F.; Petrich, J. W. J.
Phys. Chem. 1993, 97, 1770. (d) Smirnov, A. V.; English, D. S.; Rich, R.
L.; Lane, J.; Teyton, L.; Schwabacher, A. W.; Luo, S.; Thornburg, R. W.;
Petrich, J. W. J. Phys. Chem. 1997, 101B, 2758.
(7) Mente, S.; Maroncelli, M. J. Phys. Chem. A 1998, 102, 3860.
(8) Takeuchi, S.; Tahara, T. J. Phys. Chem. A 1998, 102, 7740.
(9) (a) Chachisvilis, M.; Fiebig, T.; Douhal, A.; Zewail, A. H. J. Phys.
Chem. A 1998, 102, 669. (b) Fiebig, T.; Chachisvilis, M.; Manger, M.;
Zewail, A. H.; Douhal, A.; Garcia-Ochoa, I.; de La Hoz Ayuso, A. J. Phys.
Chem. A 1999, 103, 7419.
(10) (a) Chou, P. T.; Wei, C. Y.; Wu, G. R.; Chen, W. S. J. Am. Chem.
Soc. 1999, 121, 12186. (b) Chou, P. T.; Wu, G. R.; Wei, C. Y.; Cheng, C.
C.; Chang, C. P.; Hung, F. T. J. Phys. Chem. B 1999, 103, 10 042.
(11) (a) Chou, P. T.; Wu, G. R.; Wei, C. Y.; Cheng, C. C.; Chang, C.
P.; Hung, F. T. J. Phys. Chem. B 2000, 104, 7818. (b) Chou, P. T.; Liao,
J. H.; Wei, C. Y.; Yang, C. Y.; Yu, W. S.; Chou, Y. H. J. Am. Chem. Soc.
2000, 122, 986.
(12) Mason, S. F.; Philp, J.; Smith, B. E. J. Chem. Soc. A 1968, 3051.
(13) Thistlethwaite, P. J.; Corkill, P. J. Chem. Phys. Lett. 1982, 85, 317.
(14) Thistlethwaite, P. J. Chem. Phys. Lett. 1983, 96, 509.
(15) Itoh, M.; Adachi, T.; Tokumura, K. J. Am. Chem. Soc. 1984, 106,
850.
(54) Waluk, J.; Grabowska, A.; Pakula, B.; Sepiol, J. J. Phys. Chem.
1984, 88, 1160.
(55) Pimentel, G. C., McClellan, A. L., Eds. The Hydrogen Bond; W.
H. Freeman and Co.: 1960, p 368.
(56) The pKa value was obtained through the average of data obtained
from ref 32 and ref 57.
(57) Perez, M. A. M.; Guzman, M. C. C.; Toledo, J. H.; Almeida, M.
B. J. Chem. Soc. Perkin Trans. 2 1986, 1573.
(58) These values are obtained from pKb ) pKH O - pKa where pKa
2
can be obtained from the corresponding conjugated acid of pyridinic nitrogen
(in â-CB32) and carbonyl oxygen (in ACID59).
(16) Konijnenberg, J.; Ekelmans, G. B.; Huizer, A. H.; Varma, C. A.
G. O. J. Chem. Soc. Faraday Trans. 2 1989, 85, 39.
(17) Bohra, A.; Lavin, A.; Collins, S. J. Phys. Chem. 1994, 98, 11 424.
(18) Fang, W. H. J. Am. Chem. Soc. 1998, 120, 7568.
(19) Garc´ıa-Ochoa, I.; Bisht, P. B.; Sa´nchez, F.; Martinez-Ata´z, E.;
(59) Gordon, Arnold, J.; Ford, R. A. The Chemist’s Companion; John
Wiley & Sons: New York, 1972.
(60) Gerlt, J. A.; Gassman, P. G. J. Am. Chem. Soc. 1993, 115, 11 552.
(61) Cleland, W. W.; Kreevoy, M. M. Science 1994, 264, 1887.
(62) Frey, P. A.; Whitt, S. A.; Tobin, J. B. Science 1994, 264, 1927.
Santos, L.; Tripathi, H. B.; Douhal, A. J. Phys. Chem. A 1998, 102, 8871.