Article

pubs.acs.org/JPCA

Excited-State Intramolecular Proton Transfer (ESIPT) Emission of

Hydroxyphenylimidazopyridine: Computational Study on Enhanced
and Polymorph-Dependent Luminescence in the Solid State
Yasuhiro Shigemitsu,*,†,§ Toshiki Mutai,‡ Hirohiko Houjou,‡ and Koji Araki*,‡
†
Industrial Technology Center of Nagasaki, 2-1303-8 Ikeda, Omura, Nagasaki 856-0026, Japan
‡
Institute of Industrial Science, University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan
§
Graduate School of Engineering, Nagasaki University, 1-14, Bunkyo-machi, Nagasaki 852-8521, Japan
*
S Supporting Information

ABSTRACT: Although 2-(2′-hydroxyphenyl)imidazo[1,2-a]-

pyridine (HPIP) is only weakly fluorescent in solution, two of its
crystal polymorphs in which molecules are packed as stacked pairs
and in nearly coplanar conformation exhibit bright excited-state
intramolecular proton transfer (ESIPT) luminescence of different
colors (blue-green and yellow). In order to clarify the enhanced and
polymorph-dependent luminescence of HPIP in the solid state, the
potential energy surfaces (PESs) of HPIP in the ground (S0) and
excited (S1) states were analyzed computationally by means of ab
initio quantum chemical calculations. The calculations reproduced
the experimental photophysical properties of HPIP in solution,
indicating that the coplanar keto form in the first excited (S1) state
smoothly approaches the S0/S1 conical intersection (CI) coupled
with the twisting motion of the central C−C bond. The S1−S0 energy gap of the keto form became sufficiently small at the
torsion angle of 60°, and the corresponding CI point was found at 90°. Since a minor role of the proximity effect was indicated
experimentally and theoretically, the observed emission enhancement of the HPIP crystals was ascribed to the following two
factors: (1) suppression of efficient radiationless decay via the CI by fixing the torsion angle at the nearly coplanar conformation
of the molecules in the crystals and (2) inhibition of excimer formation resulting from the lower excited level of the S1-keto state
compared to the S0−S1 excitation energy in the enol form. However, the fluorescence color difference between the two crystal
polymorphs having slightly different torsion angles was not successfully reproduced, even at the MS-CASPT2 level of theory.

■ INTRODUCTION
Luminescent organic solids have been studied extensively
because of their potential for use in various photofunctional
applications such as nonlinear optics,1−3 organic light-emitting
diodes,3−6 and fluorescent biosensors.7−9 In general, fluores-
cent organic compounds tend to lose their emissive properties
in the condensed phase owing to the various pathways of
radiationless energy dissipation through stacking and other
intermolecular interactions. Therefore, it is essential to suppress
radiationless deactivation of the excited state for compounds in
the solid state and many efforts have been devoted to this topic.
Recently, a novel class of luminescent compounds that do not
(or only weakly) show luminescence in solution but exhibit
greatly enhanced luminescence in the solid state or upon
formation of aggregates have been reported and are attracting
considerable interest.10−22 The process is referred to as
aggregation-induced emission enhancement (AIEE), but
examples of AIEE systems are still quite limited. Aromatic
siloles by Tang et al.,11 phenylenevinylenes by Park et al.,12
terpyridine by our group,23 and several other compounds have
been reported to be AIEE systems, and they all have similar
structural characteristics. In these compounds, two or more
aromatic systems are connected by a single bond, and a critical
role of the rotation around the single bond in the excited state
has been indicated. Though several discussions on the origin of
the AIEE effect have appeared, the mechanism of the packing-
to-luminescence transduction in these compounds is not
sufficiently clear.
Applications based on the tuning and switching of solid-state
luminescence are attracting considerable interest, and increas-
ing numbers of reports on the tuning and switching of solid-
state luminescence by controlling molecular packing have
recently appeared. We previously reported polymorph-depend-
ent luminescence of 2-(2′-hydroxyphenyl)imidazo[1,2-a]-
pyridine (HPIP),24 which has the common structural features
of other AIEE systems.25 HPIP shows very weak fluorescence
with a large Stoke’s shift in apolar solvent, which is ascribed to
excited-state intramolecular proton transfer (ESIPT) emission,
Received: August 27, 2012
Revised: November 17, 2012

© XXXX American Chemical Society A dx.doi.org/10.1021/jp308473j | J. Phys. Chem. A XXXX, XXX, XXX−XXX
The Journal of Physical Chemistry A
as illustrated in Figure 1. However, two crystal polymorphs of
HPIP exhibit bright luminescence of different colors, blue-green
Figure 1. (a) Tautomeric structures of HPIP and (b) normal and
ESIPT luminescence.
and yellow, both with high quantum yield, showing a clear
AIEE effect. Furthermore, switching between the luminescence
colors can be achieved by thermal interconversion of the two
polymorphs, offering an excellent example of packing-
controlled luminescence color tuning.
ESIPT luminescence has attracted much attention in various
fields of photochemistry and photophysics.26,27 In the present
study, we focused on the bright ESIPT luminescence of HPIP
in the solid state. Although the excited states associated with
the inter-ring twisting motion have been investigated from a
theoretical viewpoint,28−31 the mechanistic details remain
unclear. Herein, we report the potential energy surfaces
(PESs) of HPIP in the ground (S0) and excited (S1) states
explored by means of ab initio quantum chemical calculations.
Attempts to elucidate the origin of emission enhancement in
HPIP in the solid state and the possible mechanism for the
packing-controlled luminescence color tuning based on both
the experimental and theoretical studies are also discussed.
■ COMPUTATIONAL DETAILS
The S0 geometries of HPIP were optimized using density
functional theory (DFT) with the 6-31G(d,p) basis set in
conjunction with the B3LYP functional. For the optimized S0
geometries, the vertical excitation energies were evaluated with
time-dependent DFT (TD-DFT) at the TD(B3LYP)/6-
31+G(d,p) level. As reference, equation-of-motion with
coupled cluster singles and doubles (EOM-CCSD) were
employed to estimate the transition energies using cc-pVDZ
basis set. Alternatively, the S0 geometries were optimized with
state averaged CASSCF(6e,6o) level of theory over the lowest
four states with even weight 0.25, where 6π-electrons were
distributed into 3π (HOMO, HOMO−1, HOMO−2) and 3π*
orbitals (LUMO, LUMO+1, LUMO+2), using atomic natural
orbital small basis set (ANO-S)32 with the contractions
C,N,O[3s2p1d]/H[2s1p]. The quantitative vertical excitation
energies including dynamical electron correlation were
evaluated by multistate CASPT2 (MS-CASPT2).33 The
single-point MS-CASPT2 calculations employed 4 to 8 state-
averaged CAS (10e,9o), which include one lone pair occupied
B dx.doi.org/10.1021/jp308473j | J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
orbital localized on N3 (enol form) or one lone pair occupied
orbital localized on O1 (keto form), using atomic natural orbital
large basis set (ANO-L)34 with the contractions C,N,O-
[4s3p2d]/H[3s2p]. A real level-shift parameter35 of 0.3 was
employed for all MS-CASPT2(10,9)/ANO-L calculations with
careful monitoring of the excited state reference weights. The
S1 geometries and the energy profile as a function of the torsion
angle (ϕ) were evaluated by CASSCF(6,6)/ANO-S. The fully
relaxed S1 PESs were explored at CASSCF(6,6)/ANO-S-MB
(minimal base) level. Solvent effects were not considered
throughout these computations.
The DFT, TD-DFT, and EOM-CCSD calculations were
performed with Gaussian09.36 The CASSCF and MS-CASPT2
calculations were performed with MOLCAS 7.4.33
■ RESULTS AND DISCUSSION
Photophysical Properties of HPIP. In order to enhance
the present discussion, previously reported24 and newly
obtained photophysical properties of HPIP in solution,
together with those in the solid state, are summarized in
Table 1. HPIP shows almost identical absorption bands at
Table 1. Photophysical Properties of HPIP
λabs
(nm) λem (nm) (Φ, τ (ns))
solution ethanol 327 373 (0.11,
1.6)
cyclohexane 337 578 (0.04,
0.73)
THF 332, 346 377 (0.08, 602 (0.02,
2.6) 0.52)
THF (77 K)a 332, 349 370 (−, 2.7) 521 (−, 5.3)
solid polymer matrixb 332 523 (0.37, 4.9)
amorphous 337 537 (0.39, 6.4)
powder
BG-form crystal 334 496 (0.50, 5.9)
Y-form crystal 335 529 (0.37, 5.8)
a
Frozen state at 77 K. bIn PMMA film.
around 330−340 nm in common organic solvents but displays
two different solvent-dependent fluorescence bands, i.e, very
weak ESIPT fluorescence that is yellow to orange in color in
apolar and aprotic cyclohexane, normal blue fluorescence in
protic ethanol, and both normal and ESIPT dual fluorescence
in THF (Figure 2). According to the study by Douhal et
al.,37−40 the enol form in the planar conformation is the most
stable form in the ground state owing to stabilization of the
intramolecular hydrogen bond between the imino and phenolic
groups.
The two crystal polymorphs of HPIP, the blue-green-
emitting (BG) and yellow-emitting (Y) crystals, showed
markedly enhanced and polymorph-dependent ESIPT lumi-
nescence with high quantum yields (Table 1). The crystal
structures of BG (Pbca, Z = 8) and Y (P21/c, Z = 8) are
shown24 in Figure 3 and Supplementary Table S-1. The Y
crystal consists of two further conformers, Y1 and Y2 (Figure
3b). In the crystals, HPIP is in the enol form, and no apparent
distortion of the structure is observed. The hydrogen bond
between the phenol OH and the ring nitrogen is formed
exclusively in the intramolecular mode, and absence of an
intermolecular hydrogen bond is confirmed. In both crystals,
two HPIP molecules form a stacked pair unit in an antiparallel
mode (Figure 3c,d), and the pair units are aligned in a zigzag
The Journal of Physical Chemistry A
Figure 2. (a) Solid-state absorption and luminescence spectra of the
BG form (dotted line) and Y form (solid line). (b) Absorption and
emission of HPIP at room temperature (broken line) and at 77 K
(solid line) in THF solution. (Reproduced with permission from ref
24. Copyright 2008 Wiley.)
Figure 3. ORTEP structures of HPIP in the BG (a) and Y (b) crystals.
Top and side views of the stacked pairs of the BG (c) and Y (d)
crystals.
fashion (Supplementary Figure S-2). Therefore, the mode of
molecular packing in the two crystals is not significantly
different. However, closer examination of the structures shows
that there is a slight difference between the N3−C10−C11−C16
torsion angles (ϕ) of BG (5.88°) and the two conformers of Y
(Y1/Y2: 1.38°/−1.0°). In the crystals, the molecules are packed
closely together, and their conformations are strictly confined.
The effect of molecular packing on the luminescent properties
can be generally understood from the following two factors: (1)
packing-induced conformational fixation and/or alteration at a
single molecular level and (2) intermolecular interactions with
neighboring molecules. Since the amorphous powder of HPIP
also showed bright yellow luminescence, it is unlikely that the
specific mode of molecular packing is required for the observed
emission enhancement in the solid state. It should also be
noted that HPIP in a polymethylmethacrylate matrix (0.5 wt
%) and in a frozen dilute THF solution (1.00 × 10−5
mol·dm−3) at 77 K displayed bright yellow ESIPT lumines-
C dx.doi.org/10.1021/jp308473j | J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
cence, showing marked luminescence enhancement. Since
HPIP was surrounded by the polymer matrix or solvent
molecules in these cases, the results strongly indicate that the
major reason for the observed emission enhancement of HPIP
in the solid state should be ascribed to factor (1).
Ground State Geometry and Absorption Spectra. The
relative keto−enol stability was computationally analyzed, and
the results are shown in Table 2. The enol form (S0-enol) was
Table 2. Total and Relative Energies of Coplanar HPIP in
the Ground Statea
B3LYP/6-31G CASSCF(6,6)/ MS-CASPT2(10,9)/
species +(d,p) ANO-S ANO-L
enol −686.1873 −682.0288 −684.4210
keto (not found) −681.9744(+142) −684.3838(+97.6)
(ϕ = 0)
a
Total energies are in a.u., and energies of the keto form relative to the
enol form are in kJ/mol in parentheses.
always more stable than the keto form (S0-keto), and the most
stable conformations of S0-enol and S0-keto were both
confirmed to be the coplanar conformation. High stability of
the coplanar conformation is attributed to the formation of the
intramolecular hydrogen bond. The optimized geometries of
S0-enol and S0-keto forms in the coplanar conformation are
shown in Supplementary Figure S-1. The energy difference
between S0-keto and S0-enol states was computed to be 142
kJ·mol−1 by CASSCF(6,6)/ANO-S//B3LYP/6-31G(d,p) and
97.6 kJ·mol−1 by MS-CASPT2(10,9)/ANO-L//B3LYP/6-31G-
(d,p). Therefore, HPIP exists predominantly as the coplanar
enol form in the ground state.
Absorption spectra of HPIP in various solvents showed a
single band at around 330 nm. The S0 → S1 vertical transition
energies in the coplanar S0-enol form were estimated for several
low-lying states by means of TD-DFT(B3LYP), CASSCF(6,6)/
ANO-S, and MS-CASPT2/ANO-L as shown in Table 3.
Computation by the three methods predicted five ππ* excited
states with no interleaf of the nπ* state. The energy level of the
Franck−Condon (FC)-S2 was sufficiently separated from that
of the FC-S1 by 0.61 eV at the MS-CASPT2(10,9)/ANO-L
level. The S0 → S1 excitation energy predicted by CASSCF was
4.44 eV, a considerable overestimation of the experimental
value of 3.68 eV. The lack of dynamic electron correlation in
CASSCF tends to overestimate vertical excitation energies,
especially for ionic or charge-transfer states.41 The over-
estimated energy was lowered by MS-CASPT2 to 3.95 eV. The
S0 → S2 and higher excitation energies were also overestimated
by CASSCF, which were also adequately lowered by MS-
CASPT2. EOM-CCSD/cc-pVDZ results were comparable to
those obtained by CASSCF for S0 → S1 and to MS-CASPT2
for the higher energies. The agreements for S0 → S1 would be
improved with more extensive basis such as aug-cc-pVTZ or
higher, which go beyond our present computational capacity.
As shown in Table 4, excitation to the FC-S1 can be
dominantly described as HOMO−LUMO excitation of the
entire π system, while excitation to the FC-S2 state is mainly
described by (HOMO−1)−LUMO excitation with moderate
charge transfer from the phenyl ring to the imidazopyridine
ring, where (HOMO−1) resides on the phenyl ring and
LUMO is delocalized on the whole molecular plane. The
charge redistribution upon excitation to FC-S1 and FC-S2 did
not show large charge transfer between the phenolic and
The Journal of Physical Chemistry A Article

Table 3. Low-Lying Excited States of the Coplanar Enol Form of HPIP for the S0-Optimized Geometry (in eV)
excited state TD-B3LYP/6-31+G(d,p) EOM-CCSD/cc-pVDZ CASSCF(6,6)/ANO-S MS-CASPT2(10,9)/ANO-L experiment (in cyclohexane)
2A′(S1) 3.55 4.31 4.44 3.95 3.68
3A′(S2) 4.07 4.61 5.71 4.56
4A′(S3) 4.25 4.99 6.38 4.93
1A″(S4) 5.46 5.76 6.63 6.34
2A″(S5) 6.10 5.80 6.94 6.85

Table 4. Natural Orbital Occupations of the S0, S1, and S2 States of the Coplanar Enol Form of HPIPa

a
Occupations calculated using 8 state averaged CASSCF(10,9)/ANO-L with even weight 0.125.

Table 5. Natural Orbital Occupations of the S0, S1, and S2 States of the Coplanar Keto Form of HPIPa

a
Occupations calculated using 8 state averaged CASSCF(10,9)/ANO-L with even weight 0.125.

imidazopyridine moieties, according to Mulliken population
analysis (Supplementary Table S-1). More specifically, the
positive charge density is localized on the C7 and C10
imidazopyridine carbons, whereas the negative charge density
remained on the N3, C5, and C8 imidazopyrizine atoms. As a
result, no drastic change in the dipole moment was observed
upon the S1 or the S2 excitations. This agrees well with the
small solvent dependence observed for the absorption maxima
in solutions (Table 1). The S2 character strongly depends on
the computational level; the main configuration is expressed in
HOMO−LUMO excitation by 4 state-averaged CASSCF(6,6)/
ANO-S and in (HOMO−1)−LUMO excitation by 4 state-
averaged CASSCF(10,9) level, respectively. The present largest
scale calculation (8 state-averaged CASSCF(10,9)) may not
give the definitive answer owing to insufficient inclusion of the
active space. According to Platt’s nomenclature,42,43 the two
low-lying excited states can be approximately described with the
four-orbital model containing HOMO−1, HOMO, LUMO,
and LUMO+1. That is, the S1 and S2 states are ascribed to La
and Ba states, respectively.
The S0-keto form possesses a larger dipole moment than the
S0-enol form (Supplementary Table S-2), owing to the localized
electron distribution on the hydroxyphenyl ring in the S0-keto
form. As shown in Table 5, upon S0 → S1 (HOMO−LUMO)
and S0 → S2 HOMO−(LUMO+1) excitations, a drastic
decrease in the dipole moment was invoked by intramolecular
electron transfer from the phenol ring to the imidazopyridine
moiety, for instance, an increase in electron density on N3 and a
decrease on O1.
D dx.doi.org/10.1021/jp308473j | J. Phys. Chem. A XXXX, XXX, XXX−XXX
S1 PES Scan along the Central C−C Bond Torsion.
Historically, the theoretical interpretation of radiationless
transitions from the excited state in polyatomic molecules has
been roughly categorized into two models: the proximity effect
model and the conical intersection (CI) model. The proximity
effect model44−51 has been applied to many luminescent/
nonluminescent heterocyclic compounds sensitive to temper-
ature, chemical substitution, hydrogen bonding, and solvents.
This sensitivity has been attributed to the vibronic couplings
invoked by near-degeneracy of the nπ*/ππ* excited levels with
out-of-plane deformation of the aromatic ring in excited states.
However, the involvement of CI on PESs has been widely
employed as an alternative model to explain radiationless
decay.52−54 CIs can be accessed directly or indirectly,55−59
acting as efficient decay funnels from excited states to ground
states within a single vibrational period. A critical role of CIs
has recently been reported for many compounds including 2-
(2′-hydroxyphenyl)benzo[d][1,2,3]triazole (BZT), 2-hydroxy-
phenyl-1,3,5-triazine, and related compounds. BZT, known as
an effective UV absorbent, forms a strong intramolecular
hydrogen bond in the ground state. Paterson et al. reported a
detailed systematic quantum chemical study of BZT, showing
that fast excited energy dissipation through CI is essential for its
high photostability.60 Additionally, ab initio computational
analyses of the luminescence of HPIP-related compounds have
been recently reported using the DFT, TD-DFT, and CIS
levels of theory.61 However, the excited-state PESs have not
been satisfactorily explored using a multiconfigurational level of
theory for this series of compounds.
The Journal of Physical Chemistry A Article

Table 6. Vertical Transition Energies of Coplanar HPIP for the S1-Optimized Geometryc,d
enol keto (ϕ = 0) keto (ϕ = 180) experiment
states ΔE strength ΔE strength ΔE strength ΔEa ΔEb
2A′(S1) 3.46 0.057 2.29 0.247 2.17 0.326 3.43 2.21
3A′(S2) 4.67 0.027 3.53 0.018 3.74 0.027
4A′(S3) 5.55 0.050 4.26 0.551 4.14 0.376
1A″(S4) 5.80 0.21 × 10−2 3.74 0.27 × 10−6 3.33 0.947 × 10−6
2A″(S5) 6.74 0.55 × 10−3 4.99 0.11 × 10−5 4.69 0.185 × 10−3
3A″(S6) 7.37 0.017 5.42 (forbidden) 5.07 0.139 × 10−3
4A″(S7) 9.54 0.69 × 10−3 7.80 0.23 × 10−6 5.85 0.113 × 10−4
a
In ethanol. bIn cyclohexane. cGeometries optimized using 4 state averaged CASSCF(6,6)/ANO-S. dVertical excitation energies calculated by 4 state
MS-CASPT2(10,9)/ANO-L.

The S1 → S0 vertical transition energies of the three coplanar

forms for the S1-optimized geometries are shown in Table 6.
The emission peak observed at 3.43 eV in ethanol can be
assigned to the enol S1 → S0 emission 3.46 eV, and the peak at
2.21 eV in cyclohexane can be assigned to the keto S1 → S0
emission 2.29 eV, in quantitative accuracy within the MS-
CASPT2 level. In cyclohexane, the most stable closed-keto
form was responsible for the red-shifted emission. The emission
maximum of the keto form (ϕ = 180) was predicted at 2.17 eV,
with a small red shift from that of the keto (ϕ = 0) being 2.29
eV. The enol form was stabilized in ethanol by intra- and
intermolecular hydrogen bonds with solvents, and this was
responsible for the blue-shifted emission. In THF at room
temperature, dual fluorescence was observed because both the
keto (ϕ = 0) and enol forms can coexist in equilibrium. For all
three species, the lowest nπ* emission was more than 1 eV
higher than the S1 state.
Geometry optimization under the planar constraint for the S1
states of the enol and keto forms were performed using
CASSCF(6,6)/ANO-S. The coplanar S1 states (2A′) of the
enol form were found to be pure ππ* states with little
involvement of the nπ* states because the lowest nπ* state
(1A″) is 2.19 eV higher than the 2A′ state. The coplanar S1-
keto form was more stable (by 32.5 kJ·mol−1) than the S1-enol
form using MS-CASPT2 as shown in Table 7. This result
Table 7. Total and Relative Energies of Coplanar HPIP
Species for the Geometries Optimized in the S1 Statea
species CASSCF(10,9)/ANO-Lb MS-CASPT2(10,9)/ANO-Lc
enol −682.0276 −684.2709
keto (ϕ = 0) −682.0202 (+19.4) −684.2833 (−32.5)
keto (ϕ = 180) −682.0176 (+26.2) −684.2772 (−16.5)
a
Total energies are in a.u., and energies of the keto form relative to the
enol form are in kJ/mol in parentheses. bThe 8 state averaged
CASSCF. cThe 4 state averaged MS-CASPT2.
indicates that HPIP can smoothly undergo ESIPT from the
enol FC-S1 state. The observed lifetimes of the ESIPT excited
states were 0.52 and 5.26 ns in the THF and frozen solutions,
respectively (Table 1). Since the ESIPT process is extremely
fast, occurring within a subpicosecond time scale,62−64
tautomerization to the keto form in the S1 state is completed
immediately.
We then studied the effect of the torsion angle, ϕ, in the S1-
keto state as shown in Figure 4. An increase of the torsion angle
from 0 to 45° induced a relatively small change in the keto-S1
level with a small local minimum at around 30°, which then
decreased gradually as the torsion angle increased beyond 45°.
E dx.doi.org/10.1021/jp308473j | J. Phys. Chem. A XXXX, XXX, XXX−XXX
Figure 4. S0 and S1 potential energy profiles for the keto and enol
forms at S1-optimized geometries (CASSCF(6,6)/ANO-S).
In contrast, monotonic elevation of the S0-keto level was
observed as the torsion angle increased across the whole range
of torsion angles examined. The difference in the S0 → S1
energy gap of the keto form became sufficiently small at 60°,
and the corresponding crossing point was found at 90°. The
presence of the S0/S1 CI provided the efficient and fast decay
process of the excited state, resulting in the low quantum yield
of fluorescence in solution.
Furthermore, in order to study the whole picture of HPIP in
S1, we performed fully optimized PES explorations without
constraint of the torsion angle (CASSCF(6,6)/ANO-S-MB),
including the ESIPT process and subsequent regeneration of
the S0-enol form. The results gave a PES profile similar to that
obtained by fixing the torsion angles. The reaction path from
the FC-S1 state in the enol form leads to a nonplanar keto-S1
minimum. Energetically, there is a shallow minimum of the S1-
enol level at ϕ = 6.6°, which deviated only slightly from the
coplanar conformation. However, the energy difference from
that of the planar conformation is only 4 kJ·mol−1, and as
described earlier, ESIPT is the ultrafast process. Therefore, it is
likely that the keto form is formed immediately after excitation
to the S1 state. The twisting process is associated with the enol-
to-keto tautomerization. The local minimum of the keto-S1
state is found at the twisting angle of 37.7° and is approximately
3.1 kJ·mol−1 lower in energy than the FC-S1 level. During the
process, there is no avoided crossing between the ππ* and
rapidly descending nπ* excited states. The electronic state of
the keto-S1 minimum is a valence ππ* state, which is contrasted
with a diradical-like open-shell state of BZT.60 The overall
process is a hydrogen-atom transfer (or coupled electron plus
proton transfer), as indicated by the net charge on the
The Journal of Physical Chemistry A
transferred hydrogen, which remains unchanged during the
transfer. The intersection point between the S0 and S1 states
was found at the torsion angle of 60°, approximately 19
kJ·mol−1 downhill from the keto-S1 minimum at 37.7°. The
barrier-free landscape from the keto-S1 minimum to the S0/S1
CI allows an efficient twisting motion to proceed to the CI. The
molecule at the CI switches to the keto-S0 PES, goes downhill
to the local keto-S0 minimum at the torsion angle of 10.2°, and
then surmounts the barrier between the keto and enol forms to
finally reach the original coplanar S0-enol form.
Enhanced Solid-State Luminescence of HPIP. Though
HPIP showed very weak ESIPT luminescence in solution, it
showed bright luminescence not only in crystals, but also in
frozen dilute solutions and polymer matrixes. Therefore, the
observed AIEE effect in the solid state should be explained
primarily from the single molecular level. The CASSCF study
clearly showed the presence of the S0/S1 CI, and the efficient
radiationless decay process of the molecule proceeds via CI
coupled with the twisting motion. This is the reason for the
weak emission in solution. In the frozen solution or polymer
matrix, HPIP is thought to be fixed in the most stable coplanar
enol conformation. In the BG and Y crystals, the torsion angles
are less than 6°, and the excited states at these angles are
sufficiently separated from the CI. Therefore, fixation of the
torsion angle ϕ at or close to the coplanar conformation in the
solid state should be the origin of the observed emission
enhancement in the crystalline state. Suppression of the
efficient radiationless decay process via the S0/S1 CI allows
emissive decay from the keto-S1 to the keto-S0 state on a
nanosecond time scale. The fact that no phosphorescence was
observed in the solid state or frozen solution at 77 K indicates
that no spin−orbit coupling process was involved in the S1
state. The present quantum chemical studies indicate that the
proximity effect (nπ*−ππ* coupling in the excited states) plays
a minor role in the relaxation process. Though there are many
discussions on the enhanced solid-state luminescence because
of the importance of solid-state organic luminescent materials,
the present study indicates that the suppression of efficient
radiationless decay via CI by fixation of the torsion angle at or
near the coplanar conformation is the primary reason for the
efficient luminescence in the solid-state.
However, in order to understand the observed AIEE effect of
the HPIP crystals, further discussion is required. As shown in
Figure 3b, in the HPIP crystals, two molecules are packed close
enough to form a stacked pair in an antiparallel mode, and their
inter-ring distances are the typical stacking distance of 3.4 Å.
There is sufficient molecular overlap between the stacked pair
(Figure 3), though the extent of the overlap is slightly different
for the BG and Y crystals. In the case of aromatics having planar
molecular structures, exciton migration or excimer formation
between the stacked molecules provide efficient energy
dissipation pathways.65 However, the S0−S1 excitation energy
in the enol form is much higher than the Stokes-shifted ESIPT
emission from the S1-keto state, indicating that the S1-keto state
can be regarded as an energy trap to inhibit exciton migration.
Further, there is a large difference between the electronic
structures of the keto and the enol forms. Therefore, it is likely
that there is no strong intermolecular orbital interaction to
induce excimer formation between the stacked-pair molecules.
Indeed, as shown in Table 1, the emission energies, quantum
yields, and emission lifetimes of the HPIP crystals were not
much different from those of the amorphous solid or the frozen
dilute solution, indicating that there was no additional decay
F dx.doi.org/10.1021/jp308473j | J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
process such as exciton migration or excimer formation in the
case of the HPIP crystals. Thus, the BG and Y crystals showed
efficient luminescence similar to that of the matrix-separated
state, though the stacked molecular pair was formed in the
crystals.
Polymorph-Dependent Luminescence Color of the
HPIP Crystals. From Figure 4, it is apparent that the energy
gap between the S1-keto and S0-keto states depends greatly on
the torsion angle. The energy gap gradually decreases from the
coplanar conformation to the CI point. Unless the gap becomes
too small to induce efficient radiationless decay or the torsion
angle becomes too large to prevent the ESIPT process, an
increase in the torsion angle is expected to induce red-shift of
the ESIPT luminescence. Interestingly, however, the S1-keto
level shows a small local minimum at around 30°, and an
increase of the torsion angle does not induce monotonic
decrease of the energy gap. The torsion angles ϕ in the crystals
are small enough; however, there is a slight difference (Figure
3a), i.e., the BG crystal has the torsion angle of 5.8°, and in the
Y crystal, the torsion angles of conformers Y1 and Y2 are 1.3°
and −1.0°, respectively.
Although the torsion angle of BG is larger than those of Y,
the luminescence of BG appears at the higher energy side, as
shown in Table 8. Since the difference between the emission
Table 8. Vertical Transition Energies (in eV) of HPIP for the
BG and Y1/Y2 Crystal Geometriesa
experiment
method BG Y1/Y2 BG Y1/Y2
CASSCF(10,9)/ANO-Lb 1.84 1.94/1.95 2.58 2.34
MS-CASPT2(10,9)/ANO-Lc 2.19 2.12/2.19
EOM-CCSD/cc-pVDZ 2.42 2.41/2.42
a
Geometries partially optimized with the twist angles ϕ fixed at 5.8
(BG), 1.3(Y1), and −1.0(Y2). bThe 4 state averaged CASSCF. cThe 4
state averaged MS-CASPT2.
energies and the torsion angles of the BG and Y crystals is quite
small, 0.24 eV, we conducted higher-level CASSCF(10,9)/
ANO-L and MS-CASPT2(10,9)/ANO-L calculations in order
to qualitatively estimate the effect of the torsion angle for the
emission energies. The estimated emission energies of the
conformers at each torsion angle are listed in Table 8. The
CASSCF(10,9)/ANO-L computations apparently failed to
reproduce the observed results, predicting a higher emission
energy for Y1/Y2 owing to the lack in dynamical correlation.
Though the MS-CASPT2 calculations predicted a slightly lower
energy for Y1, there was no difference between BG and Y2. The
disappointing results primarily indicate that the larger-scale MS-
CASPT2 computations are required to correctly reproduce the
small gap between BG and Y1/Y2, including the more extended
CAS, averaged states, and basis sets. Alternatively, semi-
quantitative EOM-CCSD/cc-pVDZ computations unexpect-
edly gave the best absolute agreements with the experiments,
despite the compact basis set employed. Although the predicted
gap between BG band Y1/Y2 was negligibly small, 0.01 eV,
more extensive EOM-CC computations may be promising in
the quantitative predictions. In addition, since the estimation is
based only on the conformation of the single molecule, this
failure would be derived from the oversimplified estimation
model to reproduce the small energy difference of the
polymorph-dependent blue-green and yellow luminescence. It
may be necessary to consider the effect of the surrounding
The Journal of Physical Chemistry A
molecules, degree of conformational freedom in the crystal, and
other factors66−68 in order to explain the small emission energy
difference.
■
CONCLUSIONS
In the present study, ab initio electronic structure calculations
were conducted to investigate the distinctive ESIPT lumines-
cence enhancement of HPIP in the solid state. The calculations
reproduced the experimental photophysical properties of HPIP
in solution, and the coplanar keto form in the S1 state was
indicated to approach smoothly to the S0/S1 CI coupled with
the twisting motion of the central C−C bond. The S1−S0
energy gap of the keto form became sufficiently small at the
torsion angle of 60°, and the corresponding CI point was found
at 90°. Since a minor role of the proximity effect was indicated
experimentally and theoretically, the observed emission
enhancement of HPIP in the crystals was ascribed to the
following two factors: (1) suppression of the efficient
radialtionless decay via CI by fixation of the torsion angle at
the nearly coplanar conformation in the crystals and (2)
inhibition of excimer formation resulting from the lower excited
level of the S1-keto state compared to the S0 → S1 excitation
energy in the enol form. However, the fluorescence color
difference between the two crystal polymorphs having slightly
different torsion angles was not successfully reproduced even at
MS-CASPT2 levels of theory. The reproduction of the
experiments will require the larger scale computations
considering the extended CAS, the averaged states, and the
basis sets for future tasks.
■
ASSOCIATED CONTENT
*
S Supporting Information
Crystallographic details of BG and Y1/Y2 crystals of HPIP,
Mulliken charges and dipole moments of S0-enol (planar) and
S0-keto (planar), optimized geometries of S0-enol, S0-keto, S1-
enol, and S1-keto, and crystal packing drawings of HPIP in unit
cell. This material is available free of charge via the Internet at
http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*Phone: +81-957-52-1133. Fax: +81-957-52-1136. E-mail:
shige@tc.nagasaki.go.jp (Y.S.); araki@iis.u-tokyo.ac.jp (K.A.).
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
This study was supported in part by Grants-in-Aid for Scientific
Research (B) (No. 21350109, to K. A.) and (C) (No.
20510094, to T. M.) from the Japan Society for the Promotion
of Science (JSPS) and a grant from the Science and
Technology Agency of Japan (to Y.S.). Financial assistance
(to Y.S.) from Shiseido, Co., Ltd. is also acknowledged.
■
REFERENCES
(1) Yuan, Z.; Collings, J. C.; Taylor, N. J.; Marder, T. B.; Jardin, C.;
Halet, J.-F. J. Solid State Chem. 2000, 154, 5−12.
(2) Branger, C.; Lequan, M.; Lequan, R. M.; Large, M.; Kajzar, F.
Chem. Phys. Lett. 1997, 272, 265−270.
(3) Friend, R. H.; Gymer, R. W.; Holmes, A. B.; Burroughes, J. H.;
Marks, R. N.; Taliani, C.; Bradley, D. D. C.; dos Santos, D. A.; Brédas,
J. L.; Loegdlund, M.; Salaneck, W. R. Nature 1999, 397, 121−134.
G dx.doi.org/10.1021/jp308473j | J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
(4) Doi, H.; Kinoshita, M.; Okumoto, K.; Shirota, Y. Chem. Mater.
2003, 15, 1080−1089.
(5) Chiang, C.-L.; Wu, M.-F.; Dai, D.-C.; Wen, Y.-S.; Wang, J.-K.;
Chem, C.-T. Adv. Funct. Mater. 2005, 15, 231−238.
(6) Shinar, J. Organic Light-Emitting Devices; Springer-Verlag: New
York, 2004.
(7) Kikuchi, K.; Takakusa, H.; Nagano, T. Trends Anal. Chem. 2004,
23, 407−415.
(8) Fujikawa, Y.; Urano, Y.; Komatsu, T.; Hanaoka, K.; Kojima, H.;
Terai, T.; Inoue, H.; Nagano, T. J. Am. Chem. Soc. 2008, 130, 14533−
14543.
(9) Sameiro, M.; Goncalves, T. Chem. Rev. 2009, 109, 190−212.
(10) Luo, J. D.; Xie, Z. L.; Lam, J. W. Y.; Cheng, L.; Chen., H. Y.;
Qiu, C. F.; Kwok, H. S.; Zhan, X. W.; Liu, Y. Q.; Zhu, D. B.; Tang, B.
Z. Chem. Commun. 2001, 1740−1741.
(11) Tang, B. Z.; Zhan, X.; Yu, G.; Lee, P. P. S.; Liu, L.; Zhu, D. J.
Mater. Chem. 2001, 11, 2974−2978.
(12) An, B.-K.; Kwon, S. K.; Jung, S. D.; Park, S. Y. J. Am. Chem. Soc.
2002, 124, 14410−14415.
(13) Chung, J. W.; You, Y.; Huh, H. S.; An, B.-K.; Yoon, S.-J.; Kim, S.
H.; Lee, S. W.; Park, S. Y. J. Am. Chem. Soc. 2009, 131, 8163−8172.
(14) Horiguchi, E.; Matsumoto, S.; Funabiki, K.; Matsui, M. Bull.
Chem. Soc. Jpn. 2005, 78, 1167−1173.
(15) Park, S.-Y.; Ebihara, M.; Kubota, Y.; Funabiki, K.; Matsui, M.
Dyes Pigm. 2009, 82, 258−267.
(16) Minakata, S.; Moriwaki, S.; Inada, H.; Komatsu, M.; Kaji, H.;
Ohmori, Y.; Tsumura, M.; Namura, K. Chem. Lett. 2007, 36, 1014−
1015.
(17) Ooyama, Y.; Mamura, T.; Yoshida, K. Tetrahedron Lett. 2007,
48, 5791−5793.
(18) Yeh, H.-C.; Wu, W.-C.; Wen, Y.-S.; Dai, D.-C.; Wang, J.-K.;
Chen, C.-T. J. Org. Chem. 2004, 69, 6455−6462.
(19) Shimizu, M.; Mochida, K.; Hiyama, T. Angew. Chem., Int. Ed.
2008, 47, 9760−9764.
(20) Shimizu, M.; Takeda, Y.; Higashi, M.; Hiyama, T. Angew. Chem.,
Int. Ed. 2009, 48, 3653−3656.
(21) Zhao, C.-H.; Wakamiya, A.; Inukai, Y.; Yamaguchi, S. J. Am.
Chem. Soc. 1999, 121, 10658−10659.
(22) Wakamiya, A.; Mori, K.; Yamaguchi, S. Angew. Chem., Int. Ed.
2009, 48, 3653−3656.
(23) Mutai, T.; Sato, H.; Araki, K. Nat. Mater. 2005, 4, 685−687.
(24) Mutai, T.; Tomoda, H.; Ohkawa, T.; Yabe, Y.; Araki, K. Angew.
Chem., Int. Ed. 2008, 47, 9522−9524.
(25) Seo, J.; Kim, S.; Park, S. Y. J. Am. Chem. Soc. 2004, 126, 11154−
11155.
(26) Hu, R.; Li, S.; Zeng, Y.; Chen, J.; Wang, S.; Li, Y.; Yang, G. Phys.
Chem. Chem. Phys. 2011, 13, 2044−2051.
(27) Hsieh, C.-C.; Jiang, C.-M.; Chou, P.-T. Acc. Chem. Res. 2010, 43,
1364−1374.
(28) Lim, S.-J.; Seo, J.; Park, S. Y. J. Am. Chem. Soc. 2006, 128,
14542−14547.
(29) Sobolewski, A. L.; Domcke, W. J. Phys. Chem. A 2007, 111,
11725−111735.
(30) Sobolewski, A. L.; Domcke, W. Ab Initio Reactions and
Potential Energy Functions for Excited-State Intra- and Intermolecular
Hydrogen-Transfer Processes. In Ultrafast Hydrogen Bonding Dynamics
and Proton Transfer Processes in the Condensed Phase; Elsaesser, T.,
Bakker, H. J., Eds.; Kluwer Academic Publishers: Boston, MA, 2002;
pp 93−118.
(31) Yin, S.-H.; Liu, Y.; Zhang, W.; Guo, M.-X.; Song, P. J. Comput.
Chem. 2010, 31, 2056−2062.
(32) Pierloot, K.; Dumez, B.; Widmark, P.-O.; Roos, B. O. Theor.
Chim. Acta 1995, 90, 87−114.
(33) MOLCAS, version 7.4; see Aquilante, F.; De Vico, L.; Ferré, N.;
Ghigo, G.; Malmqvist, P.-Å; Neogrády, P.; Pedersen, T. B.; Pitonak,
M.; Reiher, M.; Roos, B. O. J. Comput. Chem. 2010, 31, 224.
(34) Widmark, P.-O.; Malqvist, P.-Å.; Roos, B. O. Theor. Chim. Acta
1990, 77, 291−306.
The Journal of Physical Chemistry A Article

(35) Roos, B. O.; Anderson, K. Chem. Phys. Lett. 1995, 245, 215− (68) Wu, Q.; Peng, Q.; Niu, Y.; Gao, X.; Shuai, Z. J. Phys. Chem. A
223. 2012, 116, 3881−3888.
(36) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; et al. Gaussian 09, revision A.02; Gaussian, Inc.:
Wallingford CT, 2009.
(37) Douhal, A.; Amat-Guerri, F.; Acufia, A. U. Angew. Chem., Int. Ed.
1997, 36, 1514−1516.
(38) Douhal, A.; Amat-Guerri, F.; Acufia, A. U. J. Phys. Chem. 1995,
99, 76−80.
(39) Guallar, A.; Moreno, M.; Lluch, J. M.; Amat-Guerri, F.; Douhal,
A. J. Phys. Chem. 1996, 100, 19789−19794.
(40) Organero, J. A.; Moreno, M.; Santos, L.; Lluch, J. M.; Douhal, A.
J. Phys. Chem. A 2000, 104, 8424−8431.
(41) Wilson, S. Electron Correlation in Molecules; Dover Publications:
Mineola, NY, 2007.
(42) Platt, J. R. J. Chem. Phys. 1949, 17, 484−495.
(43) Roos, B. O.; Andersson, K.; Fuelscher, M. P.; Malqvist, P.-Å.;
Serrano-Andrés, L.; Pierloot, K.; Merchán, M. In New Methods in
Computational Quantum Mechanics; Progogine, I., Rice, S. A., Eds.;
Advances in Chemical Physics 93; John Wiley & Sons: New York,
1996; pp 219−331.
(44) Zgierski, M. Z.; Grabowska, A. J. Chem. Phys. 2000, 112, 6329−
6337.
(45) Lim, E. C. J. Phys. Chem. 1986, 90, 6770−6777.
(46) Kaczmarek, L.; Balicki, R.; Lipkowski, J.; Borowicz, P. J. Chem.
Soc., Perkin Trans. 2 1994, 1603−1610.
(47) Legourriérec, D.; Kharlanov, V.; Brown, R. G.; Rettig, W. J.
Photochem. Photobiol., A 1998, 117, 209−216.
(48) Englman, R.; Jortner, J. J. Mol. Phys. 1970, 18, 145−164.
(49) Siebrand, W.; Zgierski, M. Z. J. Chem. Phys. 1981, 75, 1230−
1238.
(50) Kestner, N. R.; Logan, J.; Jortner, J. J. Phys. Chem. 1974, 78,
2148−2166.
(51) Ulstrup, J.; Jortner, J. J. Chem. Phys. 1975, 63, 4358−4368.
(52) Gentili, P. L.; Ortica, F.; Romani, A.; Favaro, G. J. Phys. Chem. A
2007, 111, 193−200.
(53) Domcke, W.; Yarkony, D. R.; Köppel, H., Eds. Conical
Intersections: Electronic Structure, Dynamics & Spectroscopy, Advanced
Series in Physical Chemistry; World Scientific: Singapore, 2004.
(54) Gustavsson, T.; Improta, R.; Markovitsi, D. J. Phys. Chem. Lett.
2010, 1, 2025−2030.
(55) Merchán, M.; Serrano-Andrés, L. J. Am. Chem. Soc. 2003, 125,
8108−8109.
(56) Ismail, N.; Blancafort, L.; Olivucci, M.; Kohler, B.; Robb, M. A.
J. Am. Chem. Soc. 2002, 124, 6818−6819.
(57) Blancafort, L.; Cohen, B.; Hare, P. M.; Kohler, B.; Robb, M. A. J.
Phys. Chem. A 2005, 109, 4431−4436.
(58) Zgierski, M. Z.; Fujiwara, T.; Kofron, W. G.; Lim, E. C. Phys.
Chem. Chem. Phys. 2007, 9, 3206−3209.
(59) Sobolewski, A. L.; Domcke, W.; Dedonder-Lardeux, C.; Jouvet,
C. Phys. Chem. Chem. Phys. 2002, 4, 1093−1100.
(60) Paterson, M. J.; Robb, M. A.; Blancafort, L.; De Bellis, A. J. Am.
Chem. Soc. 2004, 126, 2912−2922.
(61) Chipen, F. A. S.; Krishnamoorthy, G. J. Phys. Chem. A 2009,
113, 12063−12070.
(62) Foumier, T.; Pommeret, S.; Mialocq, J.-C.; Deflandre, A. In
Femtochemistry; De Schryver, F. C., De Feyter, S., Schweitzer, G., Eds.;
Wiley-VCH: Weinheim, Germany, 2001; pp 323−333.
(63) de Klerk, J. S.; Bader, A. N.; Ariese, F.; Gooijer, C. In Reviews in
Fluorescence; Geddes, C. D., Ed.; Springer: New York, 2009; pp 271−
298.
(64) Mitra, S.; Tamai, N. Chem. Phys. 1999, 246, 463−475.
(65) Turro, N. J. Modern Molecular Photochemistry; Benjamin/
Cummings Publishing Co., Inc.: Menlo Park, CA, 1978.
(66) Fukunaga, H.; Fedorov, D. G.; Chiba, M.; Nii, K.; Kitaura, K. J.
Phys. Chem. A 2008, 112, 10887−10894.
(67) Filhol, J.-S.; Deschamps, J.; Dutremez, S. G.; Boury, B.; Barisien,
T.; Legrand, L.; Schott, M. J. Am. Chem. Soc. 2009, 131, 6976−6988.

H dx.doi.org/10.1021/jp308473j | J. Phys. Chem. A XXXX, XXX, XXX−XXX