Photoinduced proton and charge transfer in 2-(2′-hydroxyphenyl) imidazo[4,5-b]pyridine

Article abstract of DOI:10.1021/jp311709c

This paper deals with the interplay between solvent properties and isomerism of 2-(2′-hydroxyphenyl)imidazo[4,5-b]pyridine (1), and the proton and charge-transfer processes that the different isomers undergo in the first-excited singlet state. We demonstr

Full text of DOI:10.1021/jp311709c

Article
pubs.acs.org/JPCB
Photoinduced Proton and Charge Transfer in 2‑(2′-
Hydroxyphenyl)imidazo[4,5‑b]pyridine
Alfonso Brenlla,† Manoel Veiga,§ J. Luis Perez Lustres, M. Carmen Ríos Rodríguez,*
́
Flor Rodríguez-Prieto,* and Manuel Mosquera*
́ ́
Departamento de Química Física and Centro Singular de Investigacion en Química Bioloxica e Materiais Moleculares (CIQUS),
Universidade de Santiago de Compostela, E-15782 Santiago de Compostela, Spain
ABSTRACT: This paper deals with the interplay between
solvent properties and isomerism of 2-(2′-hydroxyphenyl)-
imidazo[4,5-b]pyridine (1), and the proton and charge-transfer
processes that the different isomers undergo in the first-excited
singlet state. We demonstrate the strong influence of these
processes on the fluorescence properties of 1. We studied the
behavior of 1 in several neutral and acidified solvents, by UV−
vis absorption spectroscopy and by steady-state and time-
resolved fluorescence spectroscopy. The fluorescence of 1
showed a strong sensitivity to the environment. This behavior is the result of conformational and isomeric equilibria and the
completely different excited-state behavior of the isomers. For both neutral and cationic 1, isomers with intramolecular hydrogen
bond between the hydroxyl group and the benzimidazole N undergo an ultrafast excited-state intramolecular proton transfer
(ESIPT), yielding tautomeric species with very large Stokes shift. For both neutral and cationic 1, isomers with the OH group
hydrogen-bonded to the solvent behave as strong photoacids, dissociating in the excited state in solvents with basic character.
The pyridine nitrogen exhibits photobase character, protonating in the excited state even in some neutral solvents. An efficient
radiationless deactivation channel of several species was detected, which we attributed to a twisted intramolecular charge-transfer
(TICT) process, facilitated by deprotonation of the hydroxyl group and protonation of the pyridine nitrogen.
processes, both in ground and excited-states,27−31 due to their
1. INTRODUCTION
involvement in many chemical and biological phenomena, and
Proton- and charge-transfer processes have attracted much
energy-conversion processes. The crucial role played by
interest due to their ubiquity in chemical and biological
systems. Since the pioneering works of Forster1 and Weller,2 it
hydrogen bonding and proton transfer on intramolecular
̈
charge-transfer processes has also been recently reviewed.32
is well-known that the excited-state acid−base properties of
In this work, we elucidate the ground-state equilibria and the
excited-state proton- and charge-transfer processes of 2-(2′-
hydroxyphenyl)imidazo[4,5-b]pyridine (1) (Chart 1). This
multifunctional species belongs to a family of molecules
thoroughly studied because of their rich excited-state behavior.
In the following paragraphs, we briefly summarize the main
features regarding the ground-state equilibria and photophysical
behavior of structurally related species.
(a) A ground-state tautomeric equilibrium between the
normal form and the tautomer obtained by proton transfer
from the hydroxyl group to the imidazole N was observed for 2-
(3′-hydroxy-2′-pyridyl)benzimidazole33 and 4,5-dimethyl-2-(2′-
hydroxyphenyl)imidazole34 (the analogous species for 1 are
shown in Chart 1, designed as Ni and Ti). The tautomer
proportion depends strongly on the hydrogen-bond ability of
the solvent, attaining a maximum in water. For other related
species, only minor amounts of the ground-state tautomer were
molecules may differ strongly from those in the electronic
ground state. In these cases, electronic excitation frequently
triggers protonation (photobase behavior) or deprotonation
(photoacid behavior), the solvent molecules acting typically as
proton donors or acceptors.3−6 When both the acidic and basic
groups are present in the same molecule, an ESIPT process
from the acidic to the basic site can occur, yielding a
phototautomer.3,7−9 ESIPT processes generally take place at
ultrafast rates via preformed intramolecular hydrogen bonds.10
When the acidic and basic sites do not meet the geometrical
requirements for ESIPT, molecules can still undergo an excited-
state proton transfer assisted by species with hydrogen-bond
accepting and donating abilities, these acting as bridges
between the acidic and basic sites.11−14 Excited-state proton-
transfer dyes have been thoroughly used for fundamental
studies on the dynamics of proton transfer processes5,6,10,14−18
and have found numerous applications as UV photostabil-
izers,19 fluorescent chemosensors,20,21 and promising compo-
nents for photoswitches,22−24 and organic optoelectronic
materials.25,26
Received: November 28, 2012
Revised: December 20, 2012
Published: December 20, 2012
Extensive research efforts have been devoted over the last
years to the interplay between proton- and charge-transfer
© 2012 American Chemical Society
884
dx.doi.org/10.1021/jp311709c | J. Phys. Chem. B 2013, 117, 884−896

The Journal of Physical Chemistry B
Article
cyclic complex leads to deprotonation of the pyrrole NH group
and protonation of the pyridine N.11−13,49−53 Similar processes
were observed for structurally related molecules, as for example
1-azacarbazole11−13 and 1H-pyrazolo[3,4-b]quinoline.54 For
6,7,8,9-tetrahydro-5H-pyrido[2,3-b]indole, a single proton
transfer was observed in the dimer,55 whereas only ground-
state dimerization without excited-state proton transfer was
observed for 9H-imidazo[1,2-a]benzimidazole.56
Chart 1. Molecular Structures of the Normal Form Ni of
Compound 1 and the Tautomer Ti obtained after ESIPT and
the Methylated Derivatives 1-OMe and 1-NMe
The complex features displayed by molecules with some
similarities to 1 anticipate the difficulty of understanding its
excited-state behavior. To unravel the ground- and excited-state
behavior of 1, we investigated its UV−vis absorption and
fluorescence in various solvents at different acidities. Moreover,
we previously studied the behavior of the methylated
derivatives shown in Chart 1, 2-(2′-methoxyphenyl)imidazo-
[4,5-b]pyridine (1-OMe) and 2-(2′-hydroxyphenyl)-4-
methylimidazo[4,5-b]pyridine (1-NMe), which have a smaller
number of transferable protons and may be used as models of
selected tautomers of 1.57
detected in water (e.g., 2-(2′-hydroxyphenyl)benzimidazole
(HBI)).35
Review of the literature on 1 spectra has led to some
conflicting conclusions. Krishnamoorthy, Dogra et al. inves-
tigated the behavior of 1 in several solvents at various acidities
by means of absorption and fluorescence spectroscopy, and by
quantum mechanical calculations.48,58,59 Salman et al. studied
the influence of sample concentration on the spectra of 1 in
different solvents.60 Dual fluorescence was observed for 1 in
different solvents. Krishnamoorthy, Dogra et al. interpreted the
strongly Stokes-shifted fluorescence band of 1 as being due to
the emission of the tautomer obtained by the unimolecular
ESIPT process from the hydroxyl group to the benzimidazole
N, whereas the normal-Stokes-shifted band was assigned to the
emission of conformers lacking the intramolecular hydrogen
bond and therefore unable to undergo ESIPT.48,58 On the
contrary, Salman et al. proposed that 1 forms H-bonded dimers
in solution, that upon excitation lead to double proton transfer
in a similar way to 7-azaindole. The double fluorescence of 1
would be due, according to these authors, to the simultaneous
emission of the proton-transferred H-bonded dimers and of the
ESIPT tautomer.60
(b) The proton-transferred tautomer is greatly favored in the
first-excited singlet state owing to the well-known increased
excited-state acidity of the hydroxyl group for hydroxyaromatic
compounds.3−6 An ultrafast ESIPT process occurs after
electronic excitation through the preformed intramolecular
hydrogen bond, yielding the tautomer. Significant representa-
tives of this behavior are HBI, 2-(2′-hydroxyphenyl)-
benzothiazole (HBT), 2-(2′-hydroxyphenyl)benzoxazole
(HBO), and 2-(3′-hydroxy-2′-pyridyl)benzimidazole.3,7,33,35
(c) In proton-accepting solvents, a fraction of the normal
species may establish a hydrogen bond between the hydroxyl
group and the solvent, thus hindering ESIPT and favoring
proton transfer to the solvent. The proton dissociation was
observed for HBI and related molecules in neutral aqueous
solution and ethanol.35−37
(d) The excited-state acidity of the hydroxyl group is greatly
increased for the N-protonated form of HBI and related
species. These protonated molecules are strong photoacids,
which deprotonate in the excited state in acidic aqueous and
alcoholic solutions.34,35,37−39
The study described here was aimed specifically at
establishing (a) the nature of the species present in the ground
state in solutions of 1 in several solvents at different acidities,
(b) the characteristics of the photoinduced processes under-
gone by these species, and (c) the influence of molecular
structure on the fluorescence features of the excited species.
(e) Aromatic nitrogen heterocycles of pyridine and quinoline
type are generally known to be more basic in the lowest-excited
singlet state than in the ground state.3,4 The pyridine nitrogen
in molecules of the type 2-(2′-pyridyl)benzimidazole and 2-(3′-
hydroxy-2′-pyridyl)benzimidazole behaves as a strong photo-
base, becoming protonated even in neutral aqueous solution in
the excited state.39−41
(f) For several o-hydroxyarylbenzazoles, a temperature- and
viscosity-dependent radiationless deactivation of the tautomer
obtained after ESIPT has been detected and attributed to a
large-amplitude conformational motion.8,42−45 Fluorescence
measurements43,44 and quantum mechanical calculations46−48
on a wide family of compounds led to the proposal that this
motion is connected to an intramolecular charge migration
from the deprotonated hydroxyaryl group (donor) to the
protonated benzazole (acceptor), the process leading to a
nonfluorescent charge-transfer state.
(g) Bifunctional molecules with hydrogen-bond donating
(NH) and accepting (pyridine N) groups similar to 1 form
cyclic hydrogen-bonded dimers and hydrogen-bonded com-
plexes with a matching bifunctional hydrogen-bonded partner.
The most prominent example is 7-azaindole, as its dimer is
structurally similar to DNA base pairs. Upon excitation of the 7-
azaindole dimer or complexes, a double proton transfer in the
2. EXPERIMENTAL SECTION
Compound 1 was prepared by the double condensation of 2,3-
diaminopyridine (Acros) and salycilic acid (Aldrich). Poly-
phosphoric acid (Merck) was used as solvent and catalyzer. The
reaction mixture was heated to 160−180 °C for 2 h and the
final product was washed with water several times and
recrystallized from ethanol−water (50:50). The purity of the
product was checked by fluorescence and its structure
1
confirmed by H NMR (300 MHz, dimethyl sulfoxide-d6), δ
(ppm): 7.04 (m, 2H), 7.30 (d, 1H, J = 8.3 Hz), 7.34 (d, 1H, J =
7.9 Hz), 8.08 (b, 1H), 8.12 (d, 1H, J = 7.9 Hz), 8.40 (b, 1H).
Solutions were made up in double-distilled water treated with
KMnO4 and in spectroscopy-grade solvents (Scharlau).
Aqueous solutions always contained 25% (v/v) ethanol, due
to the low solubility of 1 in pure water. Acidity was varied with
HClO4 (Fluka, 60%) in acetonitrile (Scharlau, 99.9%) and
ethanol (Scharlau, 99.9%) and with NaOH (Fluka, 98%),
885
dx.doi.org/10.1021/jp311709c | J. Phys. Chem. B 2013, 117, 884−896

The Journal of Physical Chemistry B
Article
HClO4 and acetic acid/sodium acetate (Scharlau, 99.8%) or
ammonium perchlorate/ammonia (Fluka, 98%) buffers in
aqueous solutions. In all solvents, pHc was calculated as −log
([H+]/mol dm−3). All experiments were carried out at 25 °C
and the solutions were not degassed.
UV−vis absorption spectra were recorded in a Varian Cary
3E spectrophotometer. Fluorescence excitation and emission
spectra were recorded in a Jobin Yvon, Spex Fluoromax-2
spectrofluorometer, with correction for instrumental factors by
means of a reference photodiode and correction files supplied
by the manufacturer. Sample concentrations of ∼10−5 mol
dm−3 for absorption and ∼10−6 mol dm−3 for fluorescence were
employed. Fluorescence quantum yields were measured using
quinine sulfate (<3 × 10−5 mol dm−3) in aqueous H2SO4 (0.5
mol dm−3) as standard (Φ = 0.546).61,62 Fluorescence lifetimes
were determined by single-photon timing in an Edinburgh
Instruments LifeSpec−ps spectrometer, equipped with a laser
driver and pulsed LEDs from PicoQuant (PDL-800B, PLS 310
and PLS 340), and a cooled microchannel-plate photo-
multiplier from Hamamatsu (R-3809U-50). The equipment
time resolution is ∼50 ps after IRF reconvolution.
Model equations were fitted to the experimental data using a
nonlinear weighted least-squares algorithm. All data analyses
(equation fitting, decay traces reconvolution and principal
component global analysis) were performed using in-house
routines implemented in Matlab 7.5.0 for Windows (Math-
works, Natick, Massachusetts, USA). Taking into account all
the experimental error sources, the standard uncertainty was
estimated to be around 0.05 ns for the fluorescence decay
times, 10% for the fluorescence quantum yields and 0.10 for
ground-state pKa measurements.
Figure 1. (a) Absorption spectra of 1 in aqueous solution (with 25%
ethanol) at various pHc values between 1.5 and 11.0, [1] = 3.56 × 10−5
mol dm−3. (b) Absorption spectra of 1 in ethanol at various pHc values
between 2.0 and 6.0, [1] = 4.10 × 10−5 mol dm−3. The insets show the
experimental (symbols) and calculated (solid lines) acidity-dependent
contributions of the protonated (cp), neutral (cn), and deprotonated
(cd) species.
3. RESULTS
3.1. Absorption Spectra. The absorption spectra of 1
recorded in aqueous solution with 25% ethanol at basic, neutral,
and acidic pHc are shown in Figure 1a. The absorption
spectrum obtained under neutral conditions presented some
vibronic structure and a peak maximum at 31100 cm−1. Upon
decreasing pHc, the absorption spectrum lost its vibronic
structure and shifted to the red, the new peak being located at
29000 cm−1. The maximum molar absorption coefficient was
similar for the spectra recorded under neutral and acidic
conditions (about 2.1 × 104 mol−1 dm3 cm−1). In basic media,
the first absorption band shifted further to the red (peak
position at 27600 cm−1) and showed a slightly lower molar
absorption coefficient, 1.6 × 104 mol−1 dm3 cm−1.
The absorption spectra of 1 in ethanol at different acidities
are shown in Figure 1b. In neutral conditions, the absorption
spectrum was quite similar to that recorded in aqueous
solution, and only the peak at 33600 cm−1 presented somewhat
higher intensity in ethanol. As in water, the first-band maximum
molar absorption coefficient did not change from neutral to
acidic conditions. The spectrum recorded in acidified ethanol
was only slightly red-shifted from the spectrum in acidic water.
The influence of concentration on the absorption spectrum
of 1 in acetonitrile was investigated within the 2.5 × 10−6 to 2.6
× 10−4 mol dm−3 range. As can be seen in Figure 2, no change
was observed in the spectrum within this concentration range.
A good verification of the Beer−Lambert law for solutions of 1
in acetonitrile was observed.
Figure 2. Absorption spectra of 1 in acetonitrile at different
concentrations: [1] = 2.5 × 10−6 mol dm−3 (blue), [1] = 2.5 ×
10−5 mol dm−3 (red), and [1] = 2.6 × 10−4 mol dm−3 (green).
were almost identical. The structureless emission band peaked
at 20100 cm−1 and showed an exceptionally large Stokes shift
and a fluorescence quantum yield of 0.31 (Table 1).
The absorption and fluorescence spectra of 1 in ethanol,
trifluoroethanol, and water with 25% ethanol (panels b−d of
Figure 3) were qualitatively similar to those reported in
cyclohexane, except for the fact that in these solvents the main
fluorescence band (peaking at 20500 cm−1) obtained under
excitation at 30300 cm−1 was accompanied by a mormal-
Stokes-shifted emission at about 28000 cm−1. The intensity of
this emission band was very low in water with 25% ethanol.
The excitation spectrum measured for this emission band at
27000 cm−1 in ethanol and trifluoroethanol was slightly blue-
shifted in comparison with the spectrum recorded for the main
emission band. In all these three solvents (ethanol, trifluor-
oethanol, and water with 25% ethanol), the excitation spectrum
recorded at the maximum of the main emission band was very
similar to the absorption spectrum measured in the same
solvent. The fluorescence quantum yields were much lower
3.2. Fluorescence Spectra and Lifetimes. The absorp-
tion and fluorescence spectra of 1 in cyclohexane are shown in
Figure 3a. The fluorescence excitation and absorption spectra
886
dx.doi.org/10.1021/jp311709c | J. Phys. Chem. B 2013, 117, 884−896

The Journal of Physical Chemistry B
Article
Table 1. Fluorescence Quantum Yields of 1 in Various
a
Solvents at 25 °C
solvent
ΦF
Neutral Media
cyclohexane
acetonitrile
ethanol
0.31 (30300 cm−1
0.09 (30300 cm−1
0.18 (32790 cm−1
0.14 (30300 cm−1
)
)
)
)
trifluoroethanol
0.047 (30300 cm−1
)
water (25% EtOH)
0.03 (30300 cm−1
)
Acid Media
acetonitrile
0.20 (33330 cm−1
0.18 (28570 cm−1
0.06 (33330 cm−1
0.02 (28570 cm−1
0.15 (33330 cm−1
0.10 (28570 cm−1
)
)
)
)
)
)
ethanol
trifluoroethanol
water (25% EtOH)
∼0.005 (32260 cm−1
)
)
∼0.002 (28990 cm−1
Basic Media
water (25% EtOH)
very low
a
The excitation wavenumbers are shown in parentheses.
excitation and emission spectra depended on the monitoring
wavenumber. The fluorescence decay of 1 in acidified
acetonitrile was biexponential when exciting at 32470 cm−1,
but a monoexponential decay was obtained by exciting at 30030
cm−1 (Table 2). The fluorescence spectrum of 1 in acidified
trifluoroethanol (panel c of Figure 4) showed similar features to
those found in acidified acetonitrile. The maximum of the
emission band slightly shifted to the red and the fluorescence
quantum yield decreased (Table 1) as the excitation wave-
number decreased. A biexponential decay was observed for 1 in
acidified trifluoroethanol (Table 2).
In acidified ethanol, the fluorescence spectrum of 1 did not
depend on the monitoring wavenumber. The main emission
band, with peak maximum at 20200 cm−1, exhibited a large
Stokes shift. This band showed a great similarity with the red-
shifted band of the spectrum recorded for 1 in neutral ethanol
(see panel b of Figures 3 and 4). A low-intensity emission band
with peak at 25000 cm−1 was also observed for 1 in acidified
ethanol. This emission band clearly overlapped the absorption
spectrum. The excitation spectrum measured at the red-shifted
band did not match at all the absorption spectrum recorded in
the same conditions. It presented two main bands, the first one
at 30120 cm−1 and the second one at 33110 cm−1. This
excitation spectrum was similar to that registered in acidified
acetonitrile with emission at 26320 cm−1 and that obtained in
acidified trifluoroethanol at 25640 cm−1. In acidified ethanol,
the fluorescence decay of 1 was biexponential (cf. Table 2) and
the fluorescence quantum yield depended on the excitation
wavenumber (Table 1).
●
●
Figure 3. Absorption spectra (― ― ―) and normalized
fluorescence emission and excitation spectra of 1 in (a) cyclohexane,
νexc
30770 cm−1 (green), νem
(red), (c) trifluoroethanol, νexc
cm−1 (green), νem = 27780 cm−1 (cyan), νem
(d) water with 25% ethanol, νexc
= 30300 cm−1 (green), νem
cm−1 (red).
̃
̃ ̃
= 30300 cm−1 (green), νem = 20410 cm−1 (red), (b) ethanol, νexc
=
̃
= 27790 cm−1 (cyan), νem
= 33330 cm−1 (orange), νexc
= 20830 cm−1 (red), and
= 20410
̃
= 20410 cm−1
̃
̃
= 30300
̃
̃
̃
̃
than in cyclohexane, decreasing in the series ethanol >
acetonitrile > trifluoroethanol > water (Table 1). The yields
were independent of the excitation wavenumber in all solvents
except ethanol, where the fluorescence quantum yield was
slightly higher exciting at 32790 cm−1 (0.18) than at 30300
cm−1 (0.14). The fluorescence spectrum of 1 in acetonitrile
(results not shown) was very similar to the spectrum obtained
for 1 in water with 25% of ethanol.
The fluorescence lifetimes of 1 in various solvents are listed
in Table 2. A monoexponential decay was found in cyclohexane
and in ethanol at the red-shifted band, but biexponential at the
same band in acetonitrile and trifluoroethanol. The blue-shifted
emission band was biexponential, both in ethanol and
trifluoroethanol.
The fluorescence spectrum of 1 in acidified water with 25%
ethanol (results not shown) presented similar features as those
observed in ethanol, but the fluorescence quantum yield was
much lower (Table 1). The very low fluorescence intensity of 1
in acidic and neutral aqueous solution with 25% ethanol
hampered the measurement of fluorescence lifetimes of 1 in
this solvent.
No emission could be detected for 1 in basic water with 25%
ethanol.
The absorption and fluorescence spectra of 1 in acidified
acetonitrile are shown in panel a of Figure 4. The fluorescence
887
dx.doi.org/10.1021/jp311709c | J. Phys. Chem. B 2013, 117, 884−896

The Journal of Physical Chemistry B
Article
Table 2. Global Analysis of Fluorescence Decays Obtained for 1 in Various Solvents at 25 °C, Where Lifetimes τ, Pre-
Exponential Factors B and Associated Percentages, and Reduced χ2 Values Are Shown at the Indicated Emission Wavenumbers
a
solvent
νem
̃
/cm−1
τ1/ns {B1 (%)}
τ2/ns {B2 (%)}
χ2
Neutral Media
3.99 [Ti*−syn/T*i−anti
1.81 [Ti*−syn/T*i−anti
cyclohexane
acetonitrile
20410, 19 230
21280−19610
21 280
]
]
1.00
1.04
1.04
1.03
1.05
1.05
1.05
1.07
1.08
1.11
1.06
1.31 [Ti*−syn/T*i−anti
]
{217 (52%)}
{204 (50%)}
{213 (53%)}
{275 (48%)}
20 410
{281 (50%)}
19 610
{262 (47%)}
ethanol
27 780
1.16 (89%) [Ni1Ha*nti/Ni3Hs*yn]
1.61 [Ti*−syn/T*i−anti
0.58 (66%) [Ni1Ha*nti/Ni3Hs*yn]
0.40 [Ti*−syn/T*i−anti
0.38 (11%) [Ni1Ha*nti/Ni3Hs*yn]
20 410
]
trifluoroethanol
27 780
0.27 (34%) [Ni1Ha*nti/Ni3Hs*yn]
1.26 [Cp1Ha*nti/Cp3H*syn]
{66 (33%)}
21740, 20 830
21 740
]
{427 (67%)}
{425 (78%)}
20 830
{37 (22%)}
Acid Media
acetonitrile, pHc 3.0 (νexc
̃
= 30030 cm−1
= 32470 cm−1
)
)
25640, 24390, 23260, 22 220
2.42 [Cp1Ha*nti/Cp3H*syn]
2.46 [Cp1Ha*nti/Cp3H*syn]
{378 (87%)}
1.10
1.06
1.03
1.07
1.04
1.08
1.10
1.04
1.05
1.05
1.04
1.07
acetonitrile, pHc 3.0 (νexc
̃
27030−22220
27 030
1.28 [Ci*−anti/Ci*−syn]
{106 (13%)}
{39 (4%)}
25 640
{446 (96%)}
24 390
(100%)
23 260
(100%)
22 220
(100%)
ethanol, pHc 2.0
20 410
1.53 (96%) [T*i−anti/Ti*−syn
1.23 [Cp1Ha*nti/Cp3H*syn]
{38 (5%)}
]
0.54 (4%) [T*i−anti/T*i−syn]
2.04 [Ci*−anti/Ci*−syn]
trifluoroethanol, pHc 3.0
25640−23260
25 640
{430 (95%)}
{364 (85%)}
{284 (71%)}
24 390
{104 (15%)}
23 260
{190 (29%)}
a
When a multiexponential decay function was fitted, it is indicated at each wavenumber the pre-exponential factor B and the associated percentage of
each exponential term, and the reduced χ2 value corresponding to the individual decay. Shown in square brackets are the species to which each
lifetime was assigned. When two isomeric species appear, a more concrete assignation was not possible. The excitation wavenumber was 30030 cm−1,
except otherwise stated.
the spectral changes observed for 1 (panel b of Figure 1) and 1-
OMe57 in neutral-to-acidic ethanolic solution strongly suggests
that both cations exist also in acidic solutions of 1. This
assumption will be confirmed by the fluorescence results
discussed below in section 4.4. We designate the two
protonated species as imidazolium cation (Ci) and pyridinium
cation (Cp), the corresponding equilibrium constant between
them Kip and the microscopic acidity constants for each cation
as K1(i) and K1(p) (see Scheme 1). These constants are related
by the following equation:57
4. DISCUSSION
4.1. Ground-State Acid−Base Equilibria and Proto-
tropic Equilibrium in Acid Media. The absorption spectra of
1 at different acidities displayed in Figure 1, parts a and b,
showed similar features in ethanol and aqueous solutions. Two
ground-state equilibria within the pH range 1−11 were found.
By applying principal component global analysis63,64 to the
series of absorption spectra of 1, we obtained the spectral
components associated with the protonated (p), neutral (n),
and deprotonated (d) forms, together with their experimental
and calculated acidity-dependent spectral contributions cp, cn,
and cd (insets in Figure 1), and the acidity constants pK1 and
pK2 shown in Table 3. The pK values obtained are similar to
those reported by Krishnamoorthy and Dogra.59
In keeping with previous interpretations of the acid−base
behavior of 1 and similar species, the equilibria observed for 1
in the 1−11 pH range correspond to the protonation of the
neutral molecule in acid media and deprotonation in basic
media. The deprotonation takes place at the hydroxyl group,
the most acidic position of 1. The spectrum at basic pH
corresponds therefore to a phenolate anion (A in Scheme 1).
Regarding the protonated form, 1 possesses two basic groups,
the imidazole nitrogen and the pyridine nitrogen. A thorough
study of the ground-state equilibria of the methylated derivative
1-OMe57 led us to the conclusion that both pyridinium and
imidazolium cations coexist in equilibrium for this species,
being able to estimate the absorption and fluorescence spectra
for each cation and the equilibrium constant between them in
various solvents by fluorescence spectroscopy. The similarity of
[Cp]
[Ci]
K1(i)
Kip =
=
K1(p)
To obtain an estimation of the equilibrium constant Kip
between the cations of 1 in different solvents, we evaluated
the contributions of the imidazolium and pyridinium cations to
the absorption spectra of 1 in different solvents. To this end, we
approximated the absorption spectrum of the pyridinium cation
Cp of 1 in each solvent by the experimental absorption
spectrum of protonated 1-NMe,57 which differs from Cp only in
replacing a methyl group by a hydrogen atom at the pyridine
nitrogen. For the imidazolium cation Ci, we used the
fluorescence excitation spectrum of 1 in acidified ethanol as
estimation of its absorption spectrum, since in this solvent no
emission of the pyridinium cation is observed (see section 4.4,
below). The absorption spectra of 1 in acidified solvents were
satisfactorily reproduced as a linear combination of the spectra
just described (see Figure 5). To calculate the equilibrium
constant Kip from the spectral contributions of the imidazolium
888
dx.doi.org/10.1021/jp311709c | J. Phys. Chem. B 2013, 117, 884−896

The Journal of Physical Chemistry B
Article
Scheme 1. Acid−Base Equilibria of 1, Showing the
Imidazolium (Ci) and Pyridinium (Cp) Cations, the Neutral
Normal Form (Ni), and the Anion (A)
●
●
Figure 4. Absorption spectra (― ― ―) and normalized
fluorescence emission and excitation spectra of 1 in (a) acidified
acetonitrile, [HClO4] = 1 × 10−3 mol dm−3, νexc = 28570 cm−1
̃
(green), νexc
19230 cm−1 (red), (b) acidified ethanol at pHc 2.0, νexc
(green), νem
= 20410 cm−1 (cyan), and (c) acidified trifluoroethanol at
pHc 3.0, νexc
= 28570 cm−1 (green), νexc = 33330 cm−1 (orange), νem
25640 cm−1 (cyan), νem = 20410 cm−1 (red).
̃
= 33330 cm−1 (orange), νem
̃
= 26320 cm−1 (cyan), νem
= 30300 cm−1
̃
=
̃
̃
̃
̃
̃
=
Figure 5. Experimental (green) and calculated (o o o) absorption
spectra of 1 in acidified solvents: acetonitrile ([HClO4] = 1 × 10−3
mol dm−3), ethanol (pHc 2.0), trifluoroethanol (pHc 3.0) and water
(25% EtOH, pHc 1.0). The calculated spectra were obtained by fitting
a linear combination of the absorption spectrum of 1-NMe in the
corresponding acidified solvent, as estimation of the spectrum of the
pyridinium cation Cp (red), and the excitation spectrum of 1 in
acidified ethanol, as estimation of the spectrum of the imidazolium
cation Ci (cyan). The intensities of these spectra show the calculated
contributions.
̃
Table 3. Experimental Values of the Macroscopic Acidity
Constants (pK1 and pK2) of 1 in Various Solvents at 25 °C
and Estimated Values of the Equilibrium Constant between
Imidazolium and Pyridinium Cations (Kip) and the
Microscopic Acidity Constants of each Cation (pK1(i) and
a
pK1(p)
)
solvent
acetonitrile
pK1
pK2
Kip pK1(i) pK1(p)
α
β
order acetonitrile > trifluoroethanol > ethanol > water. The
order of magnitude of the Kip constants is similar for 1 and for
the methylated derivative 1-OMe, obtained by a different
method.57 For 1-OMe, the Kip values decrease in the order
acetonitrile > ethanol > water > trifluoroethanol, changing the
position of trifluoroethanol in comparison with 1. For both, 1-
OMe and 1, the solvent dielectric constant is not the key factor
affecting the trend of the Kip values, as acetonitrile has an
intermediate value between water and the alcohols, but the Kip
values do not follow this order. The main factor that
determines the trend of the Kip values for 1-OMe seems to
be the hydrogen-bond donor ability of the solvent.57 For 1, the
trend of the Kip values changes because of the different position
of trifluoroethanol. The difference comes probably from the
fact that the methoxy group of 1-OMe does not have the
capacity to donate a hydrogen bond like the OH group of 1,
and therefore the hydrogen-bond acceptor ability of the solvent
does not affect the Kip values for 1-OMe. The values of the
parameters representing the hydrogen-bond donor (α) and the
5.7
3.2
0.19
1.51
0.83
1.17
0.31
0
trifluoroethanol
ethanol
4.51
3.53
2.3
1.9
4.0
3.1
4.4
3.3
0.77
0.18
water (25%
EtOH)
8.91
a
Last columns list the hydrogen-bond donor (α) and the hydrogen-
bond acceptor (β) ability of the pure solvents.65
and pyridinium cations, we need to know their absorption
coefficients. As the cations cannot be isolated from one another,
this information cannot be obtained experimentally. Never-
theless, we showed in a previous paper that the molar
absorption coefficients of both cations at the maximum of the
first absorption band must be similar.57 This assumption
allowed the equilibrium constants to be estimated.
The values obtained for Kip (Table 3) show that the fraction
of pyridinium cation is greater than that of the imidazolium
cation in the investigated solvents and reaches a relative
maximum in acetonitrile. The Kip values for 1 decrease in the
889
dx.doi.org/10.1021/jp311709c | J. Phys. Chem. B 2013, 117, 884−896

The Journal of Physical Chemistry B
Article
hydrogen-bond acceptor (β) ability of the solvent are listed in
Table 3.65
Pyridyl Nitrogen, and Photodissociation at the Hydroxyl
Group. Excitation of 1 in cyclohexane at the maximum of the
first absorption band led to a single fluorescence emission band
that did not overlap the excitation spectra (see panel a of Figure
3). The strongly red-shifted emission band appeared in the
fluorescence spectra of 1 in all the solvents investigated,
accompanied in all cases but in cyclohexane by a weaker
emission at ∼28000 cm−1 (Figure 3). This last emission was
similar to the fluorescence of the methylated derivative 1-
OMe;57 in contrast, the red-shifted band of 1 centered around
20000 cm−1 did not appear in the fluorescence spectrum of 1-
OMe in any solvent. These findings, together with the vast
literature on related systems with similar fluorescence
behavior,3,7,8,33,35 suggest that the normal form Ni* of 1
undergoes an ESIPT process from the hydroxyl group to the
imidazole nitrogen to yield tautomer T*i (see Chart 1),
responsible of the red-shifted emission band. This interpreta-
tion agrees with previous experimental and theoretical studies
by Krishnamoorthy, Dogra et al. on the excited-state behavior
of 1.48,58 The ESIPT process takes place at ultrafast rate for
conformers with preformed hydrogen bonds.10 This fact
explains the observation of a monoexponential decay of the
Ti* emission band in cyclohexane in the nanosecond time scale
(Table 2).
The most stable conformation of the pyridinium cation Cp
probably involves an intramolecular hydrogen bond OH···N
(see Scheme 1). In this configuration, all nitrogens are
unavailable for accepting a hydrogen bond from the solvent.
On the contrary, the imidazolium cation Ci of 1 has the
hydroxyl hydrogen free to donate a hydrogen bond to the
solvent and the lone pair of the pyridine nitrogen available to
accept a hydrogen bond. The hydrogen-bond interactions are
therefore potentially much stronger for Ci than for Cp of 1,
both the hydrogen-bond donating and accepting capacity of the
solvent being important for its stabilization. This is probably
the main factor determining that water and ethanol have the
maximum capacity for stabilizing Ci and therefore the lowest
values of Kip for 1. Trifluoroethanol, with high hydrogen-bond
donation ability, shows a higher value of the equilibrium
constant, as the lack of hydrogen-bond accepting capacity
hampers the stabilization of the OH group of Ci.
The experimental values of pK1 and the estimated values of
Kip allow calculation of the microscopic acidity constants pK1(i)
and pK1(p) (Scheme 1).57 The values obtained (Table 3)
indicate that the pyridine nitrogen is only slightly more basic
than the imidazole nitrogen of 1. A similar result was found for
the 1-OMe derivative.57
The absolute absence of normal-Stokes-shifted emission for
1 in cyclohexane indicates that this species exists in this solvent
exclusively as a closed form with intramolecular hydrogen bond
OH···N. This conclusion is supported by the good matching of
the absorption spectrum and the excitation spectrum
monitored at the red-shifted emission band (Figure 3a). Two
isomeric normal structures of 1 present an intramolecular
hydrogen bond OH···N, Ni1Hsyn and Ni3Hanti (see Chart 2).
ESIPT in these isomers yields the syn and anti conformers of
the tautomer (T*i−syn and Ti*−anti, Chart 2). The monoexponen-
tial fluorescence decay of 1 in cyclohexane suggests that only
one emitting species is present (T*i−syn or T*i−anti), and therefore
that only one of the ground-state isomers Ni1Hsyn or Ni3Hanti
exist in significant concentration in cyclohexane solution. The
quantum mechanical calculations of Chipem and Krishnamoor-
thy indicate that Ni3Hanti is more stable than Ni1Hsyn for the
isolated molecule.48 Ni3Hanti is then probably the main ground-
state isomer of 1 present in nonpolar solvents, and T*i−anti the
emissive tautomer with a lifetime of 3.99 ns in cyclohexane.
Nevertheless, our data do not allow confirmation of this
hypothesis.
4.2. Concentration and Solvent Influence on the
Absorption Spectra: No Observable Dimerization in the
Working Concentration Range. Salman et al. report that
diluted solutions of 1 in acetonitrile do not obey Beer’s law due
to the establishment of a dimerization equilibrium.60 According
to these authors, the equilibrium is completely shifted to the
dimer in tetrahydrofuran, pentane, and ethanol, while in
acetonitrile at low concentrations some dissociation of the
dimer takes place.
We studied the influence of concentration on the absorption
spectrum of 1 in acetonitrile. Figure 2 shows three spectra
obtained in the concentration range 2.5 × 10−6 mol dm−3 to 2.6
× 10−4 mol dm−3. The spectra obtained were identical,
demonstrating a perfect verification of Beer’s law at any
wavelength. The absorption spectrum of 1 obtained in
acetonitrile was also very similar to those obtained in
cyclohexane, ethanol, and water with 25% ethanol (cf. Figure
3), both in position and intensity, only slight changes in
vibrational structure being detected. This fact indicates that the
molecular state of 1 in all these solvents is the same. This state
is most probably the monomer, as the very different polarity
and hydrogen-bond ability of these solvents would originate
strong changes in the tendency of 1 to dimerize by hydrogen
bonding.
The spectrum of 1 in trifluoroethanol (see panel c of Figure
3) was also very similar to those found in other solvents, only a
very small red-shifted absorption at ∼28000 cm−1 being
detected in trifluoroethanol which is almost undetected in the
other solvents. This red-shifted absorption is probably
originated by a hydrogen-bond interaction with this solvent,
which has very strong hydrogen-bond donor ability. A similar
behavior was found in related molecules.38,54 The absence of
this shoulder in ethanol and water, and the similarity of the
spectra of 1 in these solvents and cyclohexane, strongly support
that 1 does not dimerize at room temperature in the working
concentration range in any of the investigated solvents.
4.3. Solvent-Dependent Excited-State Behavior of 1 in
Neutral Solutions: ESIPT, Photoprotonation at the
In acetonitrile, a biexponential fluorescence decay in the
tautomer band region was obtained (Table 2), suggesting that
more than one emitting species is present. No significant
variation of the relative contributions of these lifetimes with the
emission wavenumber was observed, which indicates that the
species responsible for the different lifetimes have similar
spectra. These findings suggest that these species may be the
syn and anti isomers of the tautomer, T*i−syn and Ti*−anti
,
.
originated by ESIPT after excitation of Ni1Hsyn and Ni3Hanti
Because of the higher polarity of the Ni1Hsyn form,48 the two
forms could be in comparable proportions at equilibrium in the
polar acetonitrile solvent.
The tautomer emission band of 1 in ethanol exhibited
monoexponential decay, but the measured lifetime was different
in ethanol (1.61 ns) and cyclohexane (3.99 ns). According to
the proposed Ni1Hsyn/Ni3Hanti equilibrium, it is possible that
the solute−solvent interactions cause a different isomer to be
the main ground-state species and therefore a different
890
dx.doi.org/10.1021/jp311709c | J. Phys. Chem. B 2013, 117, 884−896

The Journal of Physical Chemistry B
Article
This suggests that one or both of the isomers Ni1Hanti and
Ni3Hsyn (Chart 2) are responsible for this emission band. This
assignment is supported by the similarity of the normal-Stokes-
shifted emission of 1 and its excitation spectrum to those of the
methylated derivative 1-OMe.57 The biexponential decay
obtained for 1 in ethanol and trifluoroethanol at 27780 cm−1
(Table 2) suggests that both Ni1H*anti and Ni3Hs*yn emit, and
hence they both must also be present in the ground state.
The relative proportions of the isomers Ni1Hanti and Ni3Hsyn
of 1 in ethanol and trifluoroethanol must be very small, as the
absorption spectrum in these solvents matches the excitation
spectrum of the red-shifted tautomer band. This means that the
conformers with intramolecular hydrogen bond OH···N
(Ni1Hsyn and Ni3Hanti) are the predominant species in the
ground state. The fact that the minor ground-state components
Ni1Hanti and Ni3Hsyn show quite intense fluorescence in
ethanol and trifluoroethanol indicates that these species have a
greater fluorescence quantum yield than the tautomer in these
solvents. The higher values of the fluorescence quantum yield
measured for 1-OMe (∼0.5)57 than for 1 (Table 1) corroborate
this interpretation.
Chart 2. Molecular Structures of the Normal N and
Tautomeric T Isomers of Neutral 1
a
The relative intensity of the normal-Stokes-shifted emission
of 1 increases from cyclohexane (undetectable) to acetonitrile
(very low), ethanol and trifluoroethanol (Figure 3). This
finding, together with previous literature reports on related
systems, indicates that the hydrogen-bond ability of the solvent
is the crucial element for the appearance of ground-state species
without the intramolecular hydrogen bond OH···N (Ni1Hanti
/
Ni3Hsyn), which become stabilized by solute−solvent hydrogen-
bonding interactions. It is then surprising that the intensity of
the normal-Stokes-shifted emission of 1 in water is very low.
We consider this result to be compelling evidence for the
existence of an extra deactivation pathway of the Ni1Ha*nti/
Ni3Hs*yn isomers of 1 in aqueous media. For various related
species, it was observed that a fraction of the molecules with the
OH group hydrogen bonded to the solvent photodissociates at
the hydroxyl group to afford the anion.35,36 The same process
may take place for Ni1Ha*nti and Ni3H*syn, but the absence of
fluorescence emission from the anion of 1 (measured in basic
media, Table 1) does not allow to confirm the formation of the
excited anion in neutral media by fluorescence techniques. If
this hypothesis is true, the low intensity of the normal-Stokes-
shifted band of 1 in water indicates that most of the Ni1Ha*nti/
Ni3Hs*yn molecules do dissociate in water. As it is known, the
photodissociation is favored in water over ethanol,35 and it is
not favored in trifluoroethanol due to the low basicity of this
solvent.57
Figure 3c shows that the band appearing in the fluorescence
spectrum of 1 in trifluoroethanol has too high intensity at
24000 cm−1 to be attributable solely to the tails of the normal
and tautomer forms emission bands. As can be seen in Figure
3a, the tautomer shows almost no fluorescence at this
wavenumber, and the spectrum of the normal form, better
observed for the methylated derivative 1-OMe,57 has a very low
intensity at the mentioned wavenumber. It seems therefore that
an additional species may contribute to the fluorescence
spectrum of 1 in trifluoroethanol. By analogy with the behavior
of 1-OMe,57 we propose that excited 1 protonates in this
solvent to yield the pyridinium cation (Cp1Ha*nti in Scheme 2,
resulting from protonation of the normal form anti conformer,
but it would be Cp3Hs*yn for protonation of the syn conformer).
This hypothesis is supported by the fluorescence spectrum of
this cation measured in acidified media (see below) and the
a
Subscripts i and p indicate the imidazole or pyridine location of a H
atom. 1H and 3H indicate the position of the imidazole H. Syn and
anti conformations refer to the relative position of the O atom and the
pyridine N
phototautomer (T*i−syn/Ti*−anti) to be the emitting species in
each of these solvents. This hypothesis is supported by
semiempirical quantum mechanical calculations, which found
that Ni3Hanti is the more stable closed isomer for the isolated
molecule, but Ni1Hsyn is the more stable one in methanol
solution.58 If this is true, Ti*−syn would be the fluorescent
tautomer in ethanol solution, with a lifetime of 1.61 ns.
Nevertheless, there is no experimental evidence to confirm this
hypothesis, as the lifetime may vary with the solvent due to
different radiationless decay rates.
The fluorescence of 1 in acetonitrile, ethanol, trifluoroetha-
nol, and water showed a weak emission band with Stokes shift
in the normal range. The intensity of this band is very low in
acetonitrile and water, and more intense in trifluoroethanol and
ethanol (cf. Figure 3). The fluorescence excitation spectrum
obtained at emission wavenumbers within this band in ethanol
and trifluoroethanol differs from the excitation spectra recorded
in the red-shifted band, showing that the two emission bands
derive from different ground-state species. Taking into account
that the species responsible for the normal-Stokes-shifted
emission does not experiment ESIPT, it must be due to an
isomer of 1 without the intramolecular hydrogen bond OH···N.
891
dx.doi.org/10.1021/jp311709c | J. Phys. Chem. B 2013, 117, 884−896

The Journal of Physical Chemistry B
Article
Scheme 2. Ground-State Isomeric Equilibrium (in Black) and Excited-State Behavior (in Red) of 1 in Different Neutral
a
Solvents
a
All the structures are shown in the anti conformation of the oxygen atom and the pyridine nitrogen, but a similar mechanism is expected for the syn
conformations.
coincidence of the fluorescence lifetime assigned to it in
acidified trifluoroethanol (1.23 ns, Table 2) with one of the
lifetimes measured at the red-shifted band in neutral
trifluoroethanol (1.26 ns). As expected, this lifetime shows
larger amplitude at higher wavenumbers of the red-shifted
band, where the pyridinium cation presents its emission
maximum (∼22000 cm−1, see next section). The very short
lifetimes of the N*i forms measured at 27780 cm−1 for 1 in
neutral trifluoroethanol (Table 2) support the existence of an
extra decay path for these species in trifluoroethanol.
alcohol molecules was found for many related structures such
as 7-azaindole and 1-azacarbazole.11−13,49−53 As we already
observed that the T*p tautomer of 1-NMe is a nonemissive
species,57 the double proton transfer process proposed for Ti*
would lead to fluorescence quenching and a diminished
fluorescence quantum yield of 1 in aqueous solution. Both
the protonation and the double-proton transfer processes may
take place for T*i in water, but as they both yield nonemitting
species, the relative presence and efficiency of these reactions
remains an open question.
The major component of the fluorescence decay measured at
the red-shifted band of 1 in trifluoroethanol (0.40 ns) shows its
maximum contribution at ∼20000 cm−1, which indicates that it
must correspond to the tautomer T*i−syn or Ti*−anti formed by
ESIPT. The lifetime of this species in trifluoroethanol is much
shorter than the values measured in cyclohexane, acetonitrile
and ethanol (Table 2). It is possible that the high acidity of this
solvent and the photobasic character of the pyridyl nitrogen
induce a partial protonation of the tautomer after the ultrafast
ESIPT, yielding the tautomeric cation TC*, which undergoes a
very fast radiationless decay (see section 4.4, below, and
Scheme 3). This process would reduce the lifetime of the Ti*
tautomer and the fluorescence quantum yield of 1 in
trifluoroethanol (Table 1).
The fluorescence quantum yield of 1 in water with 25%
ethanol is even lower than in trifluoroethanol (Table 1). In
neutral water, the photoprotonation process of the pyridine
nitrogen of Ti* could also take place, as it was found in other
pyridine derivatives.40,41 However, in this solvent with both
acid and basic character, a proton transfer process can also take
place from the protonated imidazole nitrogen of T*i to the
more basic pyridine nitrogen, the solvent molecules acting as
bridges between these two functional groups. This proton-
transfer process yields the pyridine tautomer T*P (Chart 2). A
similar excited-state double-proton transfer involving water or
4.4. Solvent-Dependent Excited-State Behavior of 1 in
Acid Solutions: ESIPT and Photoacid Dissociation. The
fluorescence spectra of 1 in different acidified solvents exhibit a
complex behavior (Figure 4). In acetonitrile and trifluoroetha-
nol, the excitation and emission spectra vary significantly with
the monitoring wavenumber, showing Stokes shifts in the
normal range. The behavior in these solvents is similar to that
observed for the O-methylated derivative 1-OMe, thoroughly
discussed by us in a previous paper.57 We showed that the
imidazolium and pyridinium cations exist in equilibrium in the
ground state, exhibiting each cation a distinct absorption and
fluorescence. The imidazolium cation Ci is responsible of the
blue-shifted absorption and emission bands, whereas the
pyridinium cation Cp displays a red-shifted absorption and
emission. The excitation and emission fluorescence spectra of 1
in acidified acetonitrile and trifluoroethanol are very similar to
those of 1-OMe,57 and the biexponential fluorescence decay
exhibited by 1 in these solvents supports the existence of two
independent emitters. The wavelength-dependence of the pre-
exponential factors allows the assignment of each decay time to
the corresponding species (Table 2).
The fluorescence spectra of 1 in acidified ethanol were very
different from those found in acetonitrile and trifluoroethanol
(Figure 4). The fluorescence excitation and emission spectra
did not depend on the monitoring wavenumber, and the
892
dx.doi.org/10.1021/jp311709c | J. Phys. Chem. B 2013, 117, 884−896

The Journal of Physical Chemistry B
Article
Scheme 3. Ground-State Tautomeric Equilibria (in Black) and Excited-State Behavior (in Red) Proposed for Cationic 1 in
a
Different Solvents
a
All the structures are shown in the anti conformation of the oxygen atom and the pyridine nitrogen, but a similar mechanism is expected for the syn
conformations.
excitation spectrum did not match at all the absorption.
Identical behavior was found in water with 25% EtOH (results
not shown), but the fluorescence quantum yield is much lower
(Table 1). The similarity between the excitation spectra in
ethanol and water and the blue-shifted excitation in acetonitrile
and trifluoroethanol, assigned to the imidazolium cation Ci,
indicates that excitation of this cation is the origin of the
observed fluorescence in ethanol and water. The emission
maximum, however, is substantially red-shifted with respect to
the spectra observed in acetonitrile and trifluoroethanol,
exhibiting a large Stokes shift. On the other hand, there is a
striking similarity between the main emission band of 1
observed in acidified (Figure 4b) and neutral (Figure 3b)
ethanol, the last one attributed to the neutral tautomer Ti*
formed by ESIPT in the neutral species. This fact indicates that
the excited imidazolium ion C*i behaves as a stronger acid in
the excited state, deprotonating in solvents able to accept the
proton (ethanol and water, see Scheme 3). In acetonitrile and
trifluoroethanol,C*i does not deprotonate due to the much
lower basicity of these solvents. This behavior is typical of
hydroxy aromatic compounds, the photoacidity of the hydroxyl
group usually enhancing when the molecule is proto-
nated.2−6,34−39 This interpretation is corroborated by the
coincidence of the main fluorescence lifetime observed in the
red-shifted band in neutral and acidified ethanol (1.61 and 1.53
ns, Table 2). The minor lifetime component (0.54 ns) observed
in acidified ethanol may be attributed to the presence of a
different T*i rotamer (Ti*−syn/Ti*−anti). This minor component is
not observed in neutral ethanol, which indicates that the Ti*
precursor exists in a single conformation [Ni3Hanti or Ni1Hsyn]
in neutral ethanol, but two precursor conformers exist in acidic
ethanol [C*i−anti and C*i−syn].
Summing up, we propose that a fraction of the C*i molecules
photodissociates at the hydroxyl group in proton-accepting
solvents to yield T*i (Scheme 3), which emits the red-shifted
fluorescence band obtained for 1 in acidified ethanol and water.
We ascribe the low-intensity emission band observed at
∼25000 cm−1 to the fluorescence of undissociated Ci*, as this
band shows a normal Stokes shift with the excitation spectrum
and coincides with the emission attributed to this species in
acetonitrile and trifluoroethanol.
As we have just discussed, the fluorescence of 1 in acidified
ethanol and water is mainly due to the emission of the neutral
tautomer Ti* produced by photodissociation of the imidazolium
cation Ci*. Nevertheless, the completely different shape of the
absorption and fluorescence excitation spectra (Figure 4b)
893
dx.doi.org/10.1021/jp311709c | J. Phys. Chem. B 2013, 117, 884−896

The Journal of Physical Chemistry B
Article
clearly indicates that another cation must be present in these
solvents in larger amount than the imidazolium ion. As we have
already discussed in section 4.1, this additional species is the
pyridinium cation Cp, which according to the estimated Kip
values in ethanol and water (Table 3) is in higher proportion
than the imidazolium cation. We must conclude, therefore, that
the pyridinium cation of 1 does not emit fluorescence in
ethanol and water. This fact is supported by the behavior of the
protonated N-methylated derivative 1-NMe, model of the
pyridinium cation. This cation emits rather intense fluorescence
in acetonitrile and trifluoroethanol, but its emission could not
be detected in ethanol and water.57 The interpretation of this
behavior was thoroughly discussed by us in a previous paper,57
and we adopt here the same view, summarized in Scheme 3.
The great acidity enhancement of the hydroxyl group in the
excited state causes the conformers of the pyridinium cation
with intermolecular hydrogen bond OH···solvent (Cp1Hanti in
Scheme 3) to dissociate after excitation, producing the neutral
pyridine tautomer Tp1Ha*nti. In contrast, the conformers with
intramolecular hydrogen bond O−H···N (Cp3Hanti in Scheme
3) will undergo an ESIPT process, yielding the tautomeric
cation TCa*nti. Both Tp1Ha*nti and TCa*nti do not exist in the
ground state and are nonemissive in the excited state due to the
fact that the dissociated hydroxyl group and protonated
pyridinium nitrogen facilitate a TICT process which results in
charge-transfer structures, (Tp1H*anti(CT) and TCa*nti(CT) in
Scheme 3) with very fast radiationless deactivation.57 An
analogous behavior was found for many similar biar-
yls.42−44,46,66−73
The theoretical study of Chipem and Krishnamoorthy
provides a strong basis for the TICT hypothesis.48 Their
results indicate that torsional rotation of the T*i tautomer of 1
to form a twisted structure is one of the radiationless channels.
The twisted structure is the result of increasing the electron
density of the imidazole ring subsystem, with concomitant
pyramidalization of this ring. The charge calculations on
heterocyclic and phenolic moieties indicate the dot-dot
electronic configuration for the perpendicular geometry in the
S1 state.48 If this is so for the neutral T*i tautomer, TICT
deactivation will be even more favorable for the protonated
TC* form, as protonation of the pyridine ring will favor the
intramolecular charge transfer from the phenolate moiety. This
process will originate the total deactivation of TC* by
radiationless routes. For the related molecule 2-(4′-N,N-
dimethylaminophenyl)imidazo[4,5-b]pyridine, an emissive
TICT state has been detected. H-bonding of the solvent with
the pyridine nitrogen or alkylation of this group greatly
enhances the formation of the TICT state.74,75 Nevertheless,
the nature of the TICT state may be different for the amino
derivative and for 1, as the TICT state of 2-(4′-N,N-
dimethylaminophenyl)imidazo[4,5-b]pyridine probably in-
volves twisting of the amino group.76
water, 0.5−0.7 in other solvents).57 Moreover, the similar
fluorescence quantum yields of 1 (Table 1) and 1-NMe
(∼0.16)57 in acidified acetonitrile and trifluoroethanol confirm
that the predominant pyridinium cation of 1 behaves in the
same way as the model derivative 1-NMe.
5. CONCLUSIONS
Our results illustrate the complex interplay of solvent
properties, solute conformations, and excited-state proton-
and charge-transfer processes of 1 in solution. These processes
have a strong influence on the fluorescence properties of 1,
which switches between fluorescent and nonfluorescent states
as a function of the environment.
The main conclusions of this work are as follows:
(1) The fluorescence of 1 in neutral media is dominated by
the emission of the T*i tautomer (Scheme 2), showing a very
large Stokes shift. The planar normal forms of 1 with
intramolecular hydrogen bond O−H···N (Ni3Hanti or Ni1Hsyn)
are the predominant ground-state species in all the investigated
solvents, from cyclohexane to water. In the first excited singlet
state, the redistribution of electronic charge brings about an
increase in the acidity of the hydroxyl group, which causes fast
ESIPT along the preformed hydrogen bond.
(2) Minor fluorescence components of 1 in neutral polar
media were observed that we attributed to the normal isomers
with intermolecular hydrogen bond OH···solvent (Ni1Hanti in
Scheme 2). Depending on the solvent, this species will
fluoresce, photodissociate in water or be protonated at the
pyridine nitrogen in trifluoroethanol. The last processes
evidence the photoacid behavior of the hydroxyl group and
the photobasic properties of the pyridine nitrogen.
(3) The pyridinium Cp and imidazolium Ci cations of 1 are
in equilibrium in acid solutions (Scheme 1). The estimated
equilibrium constants indicate that Cp is always in greater
proportion than Ci, changing from a 6:1 ratio in acetonitrile to
a 2:1 ratio in water.
(4) We observed fluorescence of C*p and Ci* in solvents with
nonbasic character. In proton-accepting solvents, only
fluorescence of the neutral Ti* tautomer is observed, formed
by photoacid dissociation of the C*i hydroxyl group (Scheme
3). We propose that the excited C*p cation is also a strong
photoacid that transfers its hydroxylic proton: (i) to the nearest
benzimidazole nitrogen if a previous intramolecular hydrogen
bond exists between them (ESIPT process yielding the
tautomeric cation TC*), or (ii) to the solvent if it is a good
proton acceptor (dissociation to form the neutral Tp*
tautomer). T*p and TC* do not fluoresce probably due to an
efficient twisted intramolecular charge-transfer (TICT) process
which induces a fast radiationless deactivation. This process is
favored by deprotonated hydroxyl group and protonated
pyridine nitrogen.
In nonbasic solvents like acetonitrile and trifluoroethanol, the
photodissociation processes of cationic 1 cannot take place and
the initially excited species emit fluorescence, increasing the
fluorescence quantum yield in these solvents with respect to the
basic ones (Table 1). On the other hand, the ESIPT process
that yields a nonemissive species can take place in all the
solvents, lowering the fluorescence quantum yield for C*p of 1
with respect to that of 1-OMe, which cannot undergo ESIPT.
As Cp is the main cation in solution, this fact explains why the
fluorescence quantum yield of 1 in all the acidified solvents
investigated (Table 1) is lower than that of 1-OMe (∼0.20 in
AUTHOR INFORMATION
Corresponding Author
*E-mail: (C.R.R.) carmen.rios@usc.es; (M.M.) manuel.
mosquera@usc.es; (F.R.-P.) flor.rodriguez.prieto@usc.es.
■
Present Addresses
†Department of Chemistry, Wayne State University, 5101 Cass
Avenue, Detroit, MI 48202.
§PicoQuant GmbH, Kekules
́
trasse 7, 12489 Berlin, Germany.
Notes
The authors declare no competing financial interest.
894
dx.doi.org/10.1021/jp311709c | J. Phys. Chem. B 2013, 117, 884−896

The Journal of Physical Chemistry B
Article
(33) Rodríguez Prieto, F.; Ríos Rodríguez, M. C.; Mosquera
ACKNOWLEDGMENTS
■
Gonzal
8666−8672.
(34) Brauer, M.; Mosquera, M.; Per
́ ́
ez, M.; Ríos Fernandez, M. A. J. Phys. Chem. 1994, 98,
This work has been supported by the Spanish Government and
the European Regional Development Fund (Grant CTQ2010-
17835) and the Xunta de Galicia (Grant CN 2012/314). A.B.,
M.V., and J.L.P.L. are thankful for a “Fundacion
Davila” fellowship, a MEC-FPU fellowship, and a “Ramon y
́
ez-Lustres, J. L.; Rodríguez-
̈
Prieto, F. J. Phys. Chem. A 1998, 102, 10736−10745.
(35) Mosquera, M.; Penedo, J. C.; Ríos Rodríguez, M. C.; Rodríguez-
Prieto, F. J. Phys. Chem. 1996, 100, 5398−5407.
(36) Ríos Rodríguez, M. C.; Rodríguez-Prieto, F.; Mosquera, M. Phys.
Chem. Chem. Phys. 1999, 1, 253−260.
́
Segundo Gil
́
́
Cajal” contract, respectively.
(37) Brenlla, A.; Rodríguez-Prieto, F.; Mosquera, M.; Ríos, M. A.;
Ríos Rodriguez, M. C. J. Phys. Chem. A 2009, 113, 56−67.
(38) Penedo, J. C.; Mosquera, M.; Rodríguez-Prieto, F. J. Phys. Chem.
A 2000, 104, 7429−7441.
REFERENCES
■
(1) Forster, T. Naturwissenschaften 1949, 36, 186−187.
̈
(2) Weller, A. Z. Elektrochem. Angew. Phys. Chem. 1952, 56, 662−668.
(3) Martynov, I. Y.; Demyashkevich, A. B.; Uzhinov, B. M.; Kuz’min,
M. G. Russ. Chem. Rev. 1977, 46, 1−15.
(39) Mosquera, M.; Ríos Rodríguez, M. C.; Rodríguez-Prieto, F. J.
Phys. Chem. A 1997, 101, 2766−2772.
(4) Arnaut, L. G.; Formosinho, S. J. J. Photochem. Photobiol., A 1993,
75, 1−20.
(40) Rodríguez Prieto, F.; Mosquera, M.; Novo, M. J. Phys. Chem.
1990, 94, 8536−8542.
(5) Tolbert, L. M.; Solntsev, K. M. Acc. Chem. Res. 2002, 35, 19−27.
(6) Agmon, N. J. Phys. Chem. A 2005, 109, 13−35.
(7) Formosinho, S. J.; Arnaut, L. G. J. Photochem. Photobiol. A 1993,
75, 21−48.
(41) Novo, M.; Mosquera, M.; Rodríguez Prieto, F. J. Phys. Chem.
1995, 99, 14726−14732.
(42) Potter, C. A. S.; Brown, R. G.; Vollmer, F.; Rettig, W. J. Chem.
Soc., Faraday Trans. 1994, 90, 59−67.
(8) Ormson, S. M.; Brown, R. G. Prog. React. Kinet. 1994, 19, 45−91.
(9) Uzhinov, B. M.; Khimich, M. N. Russ. Chem. Rev. 2011, 80, 553−
577.
(10) Elsaesser, T.; Bakker, H. J. Ultrafast Hydrogen Bonding Dynamics
and Proton Transfer Processes in the Condensed Phase; Kluwer Acad.
Pub.: Dordrecht, The Netherlands, 2002.
́
(43) Ríos Vazquez, S.; Ríos Rodríguez, M. C.; Mosquera, M.;
Rodríguez-Prieto, F. J. Phys. Chem. A 2007, 111, 1814−1826.
(44) Ríos Vazquez, S.; Ríos Rodríguez, M. C.; Mosquera, M.;
́
Rodríguez-Prieto, F. J. Phys. Chem. A 2008, 112, 376−387.
(45) Barbatti, M.; Aquino, A. J. A.; Lischka, H.; Schriever, C.;
Lochbrunner, S.; Riedle, E. Phys. Chem. Chem. Phys. 2009, 11, 1406−
1415.
(11) Waluk, J. Conformational Analysis of Molecules in Excited States;
Wiley-VCH: New York, 2000.
(46) Estevez, C. M.; Bach, R. D.; Hass, K. C.; Schneider, W. F. J. Am.
(12) Chou, P.-T. J. Chin. Chem. Soc. 2001, 48, 651−682.
(13) Waluk, J. Acc. Chem. Res. 2003, 36, 832−838.
(14) Park, S.-Y.; Jang, D.-J. J. Am. Chem. Soc. 2010, 132, 297−302.
Chem. Soc. 1997, 119, 5445−5446.
(47) Tsai, H.-H. G.; Sun, H.-L. S.; Tan, C.-J. J. Phys. Chem. A 2010,
114, 4065−4079.
́
(15) Perez-Lustres, J. L.; Rodríguez-Prieto, F.; Mosquera, M.;
Senyushkina, T. A.; Ernsting, N. P.; Kovalenko, S. A. J. Am. Chem.
Soc. 2007, 129, 5408−5418.
(48) Chipem, F. A. S.; Krishnamoorthy, G. J. Phys. Chem. A 2009,
113, 12063−12070.
(16) Mohammed, O. F.; Pines, D.; Dreyer, J.; Pines, E.; Nibbering, E.
T. J. Science 2005, 310, 83−86.
(49) Taylor, C. A.; El-Bayoumi, M. A.; Kasha, M. Proc. Natl. Acad. Sci.
U.S.A. 1969, 63, 253−260.
(17) Park, S.-Y.; Kim, Y.; Lee, J. Y.; Jang, D.-J. J. Phys. Chem. B 2012,
116, 10915−10921.
(50) Avouris, P.; Yang, L. L.; El-Bayoumi, M. A. Photochem. Photobiol.
1976, 24, 211−216.
(18) Liu, W. M.; Han, F. Y.; Smith, C.; Fang, C. J. Phys. Chem. B
2012, 116, 10535−10550.
(51) Sekiya, H.; Sakota, K. J. Photochem. Photobiol. C 2008, 9, 81−91.
(52) Kwon, O. H.; Jang, D. J. J. Phys. Chem. B 2005, 109, 20479−
20484.
(19) Fluegge, A. P.; Waiblinger, F.; Stein, M.; Keck, J.; Kramer, H. E.
A.; Fischer, P.; Wood, M. G.; DeBellis, A. D.; Ravichandran, R.;
Leppard, D. J. Phys. Chem. A 2007, 111, 9733−9744.
(20) Wu, J.; Liu, W.; Ge, J.; Zhang, H.; Wang, P. Chem. Soc. Rev.
2011, 40, 3483−3495.
(53) Park, S.-Y.; Jeong, H.; Jang, D.-J. J. Phys. Chem. B 2011, 115,
6023−6031.
(54) de Klerk, J. S.; Bader, A. N.; Zapotoczny, S.; Sterzel, M.; Pilch,
M.; Danel, A.; Gooijer, C.; Ariese, F. J. Phys. Chem. A 2009, 113,
5273−5279.
(21) Zhao, J.; Ji, S.; Chen, Y.; Guo, H.; Yang, P. Phys. Chem. Chem.
Phys. 2012, 14, 8803−8817.
(55) Catalan
(56) Catalan
Mas, T. Chem. Phys. 2004, 305, 175−185.
́
, J. J. Phys. Chem. A 2010, 114, 811−816.
(22) Lim, S. J.; Seo, J.; Park, S. Y. J. Am. Chem. Soc. 2006, 128,
14542−14547.
́
, J.; de Paz, J. L. G.; del Valle, J. C.; Claramunt, R. M.;
(23) Mutai, T.; Tomoda, H.; Ohkawa, T.; Yabe, Y.; Araki, K. Angew.
Chem., Int. Ed. 2008, 47, 9522−9524.
(57) Brenlla, A.; Veiga, M.; Ríos Rodríguez, M. C.; Mosquera, M.;
Rodríguez-Prieto, F. Photochem. Photobiol. Sci. 2011, 10, 1622−1636.
(58) Krishnamoorthy, G.; Dogra, S. K. J. Lumin. 2001, 92, 91−102.
(59) Krishnamoorthy, G.; Dogra, S. K. J. Lumin. 2001, 92, 103−114.
(60) Salman, H.; Meltzman, S.; Speiser, S.; Eichen, Y. J. Lumin. 2003,
102, 261−266.
(24) Lapinski, L.; Nowak, M. J.; Nowacki, J.; Rode, M. F.;
Sobolewski, A. L. ChemPhysChem 2009, 10, 2290−2295.
(25) Park, S.; Seo, J.; Kim, S. H.; Park, S. Y. Adv. Funct. Mater. 2008,
18, 726−731.
(26) Kwon, J. E.; Park, S. Y. Adv. Mater. 2011, 23, 3615−3642.
(27) Hammes-Schiffer, S. Acc. Chem. Res. 2009, 42, 1881−1889.
(28) Hsieh, C. C.; Jiang, C. M.; Chou, P. T. Acc. Chem. Res. 2010, 43,
1364−1374.
(29) Hammes-Schiffer, S.; Stuchebrukhov, A. A. Chem. Rev. 2010,
110, 6939−6960.
(30) Weinberg, D. R.; Gagliardi, C. J.; Hull, J. F.; Murphy, C. F.;
Kent, C. A.; Westlake, B. C.; Paul, A.; Ess, D. H.; McCafferty, D. G.;
Meyer, T. J. Chem. Rev. 2012, 112, 4016−4093.
(31) Auer, B.; Soudackov, A. V.; Hammes-Schiffer, S. J. Phys. Chem. B
2012, 116, 7695−7708.
(32) Chipem, F. A. S.; Mishra, A.; Krishnamoorthy, G. Phys. Chem.
Chem. Phys. 2012, 14, 8775−8790.
(61) Melhuish, W. H. J. Phys. Chem. 1961, 65, 229−235.
(62) Crosby, G. A.; Demas, J. N. J. Phys. Chem. 1971, 75, 991−1024.
(63) Al-Soufi, W.; Novo, M.; Mosquera, M. Appl. Spectrosc. 2001, 55,
630−636.
(64) Al-Soufi, W.; Novo, M.; Mosquera, M.; Rodríguez-Prieto, F. In
Reviews in Fluorescence 2009; Geddes, C. D., Ed.; Springer: New York,
2011; p 23−45.
(65) Kamlet, M. J.; Abboud, J. L. M.; Abraham, M. H.; Taft, R. W. J.
Org. Chem. 1983, 48, 2877−2887.
(66) Maliakal, A.; Lem, G.; Turro, N. J.; Ravichandran, R.;
Suhadolnik, J. C.; DeBellis, A. D.; Wood, M. G.; Lau, J. J. Phys.
Chem. A 2002, 106, 7680−7689.
895
dx.doi.org/10.1021/jp311709c | J. Phys. Chem. B 2013, 117, 884−896

The Journal of Physical Chemistry B
Article
(67) Malval, J.-P.; Chaumeil, H.; Rettig, W.; Kharlanov, V.; Diemer,
V.; Ay, E.; Morlet-Savary, F.; Poizat, O. Phys. Chem. Chem. Phys. 2012,
14, 562−574.
(68) Paterson, M. J.; Robb, M. A.; Blancafort, L.; DeBellis, A. D. J.
Am. Chem. Soc. 2004, 126, 2912−2922.
(69) Park, S.-Y.; Jeong, H.; Yu, H.; Park, S. Y.; Jang, D.-J. Photochem.
Photobiol. 2010, 86, 1197−1201.
(70) Stsiapura, V. I.; Maskevich, A. A.; Tikhomirov, S. A.; Buganov,
O. V. J. Phys. Chem. A 2010, 114, 8345−8350.
(71) Amdursky, N.; Erez, Y.; Huppert, D. Acc. Chem. Res. 2012, 45,
1548−1557.
(72) Kovalenko, S. A.; Eilers-Konig, N.; Senyushkina, T. A.; Ernsting,
N. P. J. Phys. Chem. A 2001, 105, 4834−4843.
(73) Duvanel, G.; Grilj, J.; Chaumeil, H.; Jacques, P.; Vauthey, E.
Photochem. Photobiol. Sci. 2010, 9, 908−915.
(74) Dash, N.; Chipem, F. A. S.; Swaminathan, R.; Krishnamoorthy,
G. Chem. Phys. Lett. 2008, 460, 119−124.
(75) Dash, N.; Krishnamoorthy, G. Photochem. Photobiol. Sci. 2011,
10, 939−946.
(76) Yin, S. H.; Liu, Y. F.; Zhang, W.; Gu, M. X.; Song, P. J. Comput.
Chem. 2010, 31, 2056−2062.
896
dx.doi.org/10.1021/jp311709c | J. Phys. Chem. B 2013, 117, 884−896