Gas phase dihydrogen bonding: Clusters of borane-amines with phenol and aniline

Article abstract of DOI:10.1016/S0301-0104(02)00529-3

In this article, we present the electronic and the vibrational spectroscopy of dihydrogen bonded clusters of phenol and aniline with borane-amines in supersonic jets. Laser-induced fluorescence excitation spectroscopy was used to characterize the formatio

Full text of DOI:10.1016/S0301-0104(02)00529-3

Chemical Physics 283 (2002) 193–207
www.elsevier.com/locate/chemphys
Gas phase dihydrogen bonding: clusters of borane-amines
with phenol and aniline
G. Naresh Patwari*, Takayuki Ebata, Naohiko Mikami
Department of Chemistry, Graduate School of Science, Tohoku University, Sendai 980-8578, Japan
Received 15 February 2002
Abstract
In this article, we present the electronic and the vibrational spectroscopy of dihydrogen bonded clusters of phenol
and aniline with borane-amines in supersonic jets. Laser-induced fluorescence excitation spectroscopy was used to
characterize the formation of binary clusters. Fluorescence-detected infrared spectroscopy was used to probe the vi-
brations in the hydride (O–H and N–H) stretching region. The spectral shifts in the electronic as well as vibrational
transitions upon cluster formation indicate that phenol and aniline are hydrogen bonded to borane-amines. Density
functional theory-based calculations were carried out to determine the structures of the binary clusters. The formation
of B–Hꢀ ꢀ ꢀH–X (X ¼ O, N) dihydrogen bonded clusters between borane-amines and phenol/aniline was deduced by
comparing experimental and theoretical results. Comparison of the present results with other hydrogen bonded clusters
of phenol and aniline indicates that dihydrogen bonds are comparable in strength to conventional r-type hydrogen
bonds. Ó 2002 Elsevier Science B.V. All rights reserved.
1. Introduction
hancing the selective binding of substrates to these
biological macromolecules and as a precursor to
proton transfer reactions [1]. Although not nu-
merous, a reasonable number of reports exist in
the literature wherein the hydrogen bonding is
between an acidic proton and p-bond electrons.
Benzene–H2O is the most celebrated example of
such a p-hydrogen bonding [2,3]. The p-hydrogen
bonds are known to be rather weak compared to
those involving lone pair electrons (r-hydrogen
bonds), with typical stabilization energies roughly
half of the later. If these results are extrapolated,
straightforwardly, one could infer that r-chemical
bond would be an extremely weak hydrogen bond
acceptor. Surprisingly, however, it has been
observed in the outer coordination sphere of
Hydrogen bonding is ubiquitous in nature and
has profound impact on structure and reactivity.
The majority of hydrogen bonding observed, both
in chemistry and biology, is the interaction be-
tween an acidic proton, such as O–H or N–H
group, and a lone pair of electrons on an electro-
negative element, such as O, N or halogen. The
role of hydrogen bonding is well established in the
stabilization of biological macromolecules, en-
* Corresponding author.
E-mail addresses: naresh@qclhp.chem.tohoku.ac.jp (G.N.
Patwari), nmikami@qclhp.chem.tohoku.ac.jp (N. Mikami).
0301-0104/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved.
PII: S0301-0104(02)00529-3

194
G.N. Patwari et al. / Chemical Physics 283 (2002) 193–207
numerous transition metal complexes, that the
metal–hydrogen (M–H) r-bonds act as proton
acceptors to acidic groups, such as N–H or O–H,
forming an unconventional hydrogen bond of the
type M–Hꢀ ꢀ ꢀH–X (X ¼ N, O), termed as ‘‘dihy-
drogen bond’’ [4,5].
From the large set of crystallographic data
available for the dihydrogen bonds both in tran-
sition metal complexes and in borane-amines, the
dihydrogen bond can be recognized as an electro-
static/dipolar interaction between two oppositely
charged hydrogen atoms, of the type E–Hꢀ ꢀ ꢀH–X,
wherein E and X are electropositive and electro-
negative atoms, respectively, with respect to hy-
drogen. Structurally, the dihydrogen bond has a
The electron deficiency of borane ðBH3Þ is well
known and is the primary reason for its rich and
exotic chemistry [6]. Borane is often stabilized by
electron donation from the lone pair of electrons
on various amines resulting in a class of com-
pounds called as borane-amines, which are re-
markably stable. The simplest of borane-amines
is the borane-ammonia adduct ðBH3–NH3Þ,
which is isoelectronic with ethane ðCH3–CH3Þ.
However, the physical properties of these two
isoelectronic molecules are drastically different.
Borane-ammonia is a solid at room temperature
and has a very high melting point of +104 °C,
while ethane is a gas at room temperature, with a
melting point of )181 °C. The difference of nearly
280 °C in the melting points of these two iso-
electronic molecules cannot be explained by the
dipole moment of 5.2 D for BH3–NH3 alone.
Though the crystal structure of BH3–NH3 was
first reported in 1956 [7], only recent neutron
diffraction studies revealed the formation of the
dihydrogen bonding of the type N–Hꢀ ꢀ ꢀH–B,
between an acidic N–H and a basic B–H groups
[8]. The dihydrogen bonding forms an extended
network in BH3–NH3, thereby raising its melting
point, extensively. An earlier density functional
theory (DFT)-based calculation at PCI-80/
B3LYP level on BH3–NH3 dimer revealed the
formation of two identical dihydrogen bonds
between amine proton and borane hydride, when
two BH3–NH3 molecules are aligned in a head-
to-tail fashion [9]. The calculated interaction en-
ergy for the dimer was as large as 12.1 kcal/mol,
corresponding to a little over 6 kcal/mol for each
dihydrogen bond. This clearly indicates that the
strength of the dihydrogen bond is larger than
that of p-hydrogen bonds and is in the same
energy range as conventional r-hydrogen bonds.
Apart from BH3–NH3, dihydrogen bonding was
observed in crystals of several borane-amines re-
ported in the Cambridge Crystallographic Data-
base (CSD) [10].
ꢀ
Hꢀ ꢀ ꢀH contact distance in the range of 1.7–2.2 A.
In the case of borane-amines, the crystallographic
data show remarkable statistics of strongly bent
B–Hꢀ ꢀ ꢀH angle, averaging around 100° and near
linear N–Hꢀ ꢀ ꢀH angle, averaging around 150°. The
N–Hꢀ ꢀ ꢀH angle is tending towards linearity com-
pared to B–Hꢀ ꢀ ꢀH angle, however, is still far from
almost linearity of conventional hydrogen bonds
[9,10]. Apart from the crystallographic data, a
sizable number of reports, of quantum mechanical
calculations on dihydrogen bonding exist in the
literature [11–14]. Kulkarni [11] reported the for-
mation of a dihydrogen bond in complexes be-
tween conventional proton donors, such as
H2O=NH3, and hydrides of boron/aluminum.
Though the existence of dihydrogen bonding is
known in crystals and also has been examined
theoretically for more than a decade, only recently,
we have reported dihydrogen bonded complexes in
gas phase [15–18].
Another important aspect of the dihydrogen
bonding is the elimination of two interacting hy-
drogens in a dihydrogen bond as molecular hy-
drogen, leading to a dehydrogenation reaction.
This reaction is thermodynamically favorable due
to a large free energy (ꢁ4.3 eV) for the formation
of H2. Liu and Hoffmann [19] investigated the
formation of dihydrogen bonded LiHꢀ ꢀ ꢀHF com-
plex, which, however, is not stable owing to the
large exothermicity of the reaction LiH þ HF !
LiF þ H2. Atwood et al. [20] reported the structure
of 2,2,6,6-tetramethylpiperidine–alane complex
containing a dihydrogen bond, which, further,
readily undergoes loss of molecular hydrogen. In
this case, the dihydrogen bond can be viewed as an
intermediate for the dehydrogenation reaction.
Later an ab initio study on many group model
complexes has confirmed that the dihydrogen
bond is indeed involved in an intermediate of

G.N. Patwari et al. / Chemical Physics 283 (2002) 193–207
195
dehydrogenation reactions [21]. Recently, we have
also reported the dehydrogenation reaction from a
dihydrogen bonded cluster in the gas phase fol-
lowing resonance-enhanced multiphoton ioniza-
tion (REMPI) [22].
Our interest in dihydrogen bonding is due to its
novelty, unusual structure and reactivity. Con-
ventional hydrogen bonding is responsible for
structural stability of biological macromolecules,
such as proteins and DNA, their recognition of
substrates and in the catalysis of enzymes. Ex-
tending the analogy, it might be possible to use
dihydrogen bonding effectively in molecular rec-
ognition and catalysis. The crystal engineering of
drogen bonding was due to the fact that these two
molecules have an accessible electronic transition
in the ultra-violet region, which can be produced
conveniently using present generation tunable dye
lasers. Moreover, aniline is the simplest aromatic
amine, therefore the dihydrogen bonding between
aniline and borane-amines is nearly equivalent to
that observed in the crystals of borane-amines, and
thus can enhance our understanding of those sys-
tems. Furthermore, the spectroscopy of phenol,
aniline and their hydrogen bonded complexes is
well understood. This enables the comparison be-
tween the present dihydrogen bonded complexes
and the conventional hydrogen bonded complexes,
straightforwardly. The presentation of the paper is
as follows. In Section 2, we briefly describe the
experimental and theoretical techniques used for
the current work. In Section 3, we take up clusters
of phenol, followed by clusters of aniline and fi-
nally compare between present dihydrogen bonded
complexes and hydrogen bonded complexes. In
Section 4 we summarize and suggest the possible
future.
½ðGaH2NH2Þ Š is an example of molecular rec-
3 2
ognition using dihydrogen bonds [23]. Our moti-
vation for spectroscopic investigation of
dihydrogen bonded clusters in supersonic jets is
twofold. The first aim is the evaluation of dihy-
drogen bonding interaction energy, which is pre-
dicted by many theoretical calculations to be as
high as conventional ðrÞ hydrogen bonded com-
plexes. It would be straightforward if one observes
the characteristic shift in the hydride (O–H/N–H)
stretching vibrations of proton donating part of a
dihydrogen bond [24]. This is due to the fact that,
these stretching frequencies are very sensitive to
hydrogen bonded structures and are widely used in
the spectroscopic characterization of hydrogen
bonded complexes [25]. This would be best possi-
ble at high precision and clarity in the cold isolated
gas phase conditions provided by supersonic jets.
Moreover, only gas phase measurements can be
directly compared with the ab initio calculations
and their performance can be ascertained. Further,
it is well known that hydrogen bonding is involved
in proton transfer reactions and exchange of
r-bonds. In a similar way, the possibility of
investigating the r-bond metathesis (exchange
of two r-bonds) via dihydrogen bonding provides
another motivation for the investigation of dihy-
drogen bonded clusters in supersonic jets.
2. Experimental and theoretical techniques
Details of the experimental setup used in this
work can be found elsewhere [26]. Briefly, laser-
induced fluorescence (LIF) excitation, fluores-
cence- detected infrared (FDIR) and IR–UV
hole-burning spectra were recorded on a jet-cooled
mixture of phenol or aniline and BTMA or
BDMA. For this purpose, He at 4 bar was passed
as seed gas over phenol/aniline (Aldrich) kept in a
solvent holder at room temperature. This doped
seed gas was then mixed with vapors of either
BTMA (Kanto Chemicals) or BDMA (Acros Or-
ganics) in a sample holder heated to 315 K and
expanded using a pulsed nozzle of 0.8 mm diam-
eter. A LIF excitation spectrum of the mixture was
recorded by probing the electronic transition of
the clusters with frequency doubled output of a
Nd:YAG laser (Spectra Physics INDI-50) pumped
dye laser (Lumonics HD-500), and monitoring the
total fluorescence with a photomultiplier tube
(Hamamatsu, 1P28) combined with a cut off filter
(Corning 7–54).
In this article we report the dihydrogen bonded
clusters of phenol and aniline with borane-tri-
methylamine (BTMA) and borane-dimethylamine
(BDMA) investigated using electronic and vibra-
tional spectroscopy in supersonic jets. The choice
of phenol and aniline as proton donors for dihy-

196
G.N. Patwari et al. / Chemical Physics 283 (2002) 193–207
Infrared spectra of bare molecules and of the
inclusion of polarization and diffuse functions on
heavy atoms improves the description of hydrogen
bonding. Since dihydrogen bonding is essentially
hydrogen bonding between a pair of hydrogen
atoms, polarization and diffuse functions were in-
cluded on hydrogens as well. Moreover, the usage
of this level of theory is due to its acceptable ac-
curacy for the dihydrogen bonded structures as
exemplified in the case of BH3–NH3 and
BH3–H2O dimers [11,29] at a reasonable compu-
tational cost. Geometry optimizations for the bare
molecules and the clusters were carried out under
tight binding conditions with no constraints. The
nature of the stationary point obtained was con-
firmed by calculating the vibrational frequencies at
the same level of theory. All the structures re-
ported herein have only real frequencies. The cal-
culated vibrational frequencies were scaled by a
factor of 0.955 to correct for the basis set trunca-
tion, partial neglect of the electron correlation and
harmonic approximation, and were then compared
with the experimentally observed values. Taking
the O–H stretching vibration of bare phenol as
standard, we used the value of 0.955. The agree-
ment between the calculated and observed vibra-
tional frequencies served as a benchmark for the
structural assignment of the clusters. The stabili-
zation energies of the clusters were corrected for
the zero point energy (ZPE) and the basis set su-
perposition error (BSSE). The typical BSSE ran-
ged from 0.3 to 0.5 kcal/mol for the present
complexes. According to Kim et al. [30] 100% of
BSSE correction often underestimates the inter-
action energy and 50% correction is a good em-
pirical approximation. Therefore, we report the
stabilization energies with both 100% and 50%
BSSE corrections.
clusters were obtained by IR–UV double reso-
nance spectroscopy in combination with fluores-
cence detection, the so-called FDIR spectroscopy.
In this spectroscopic technique, the UV laser fre-
quency is fixed on an electronic transition of the
molecule/cluster of interest, and the resulting
fluorescence signal that reflects the ground-state
population of the cluster in the jet is continuously
monitored. A tunable infrared pulse is introduced
prior to the UV laser, which when resonant with a
vibrational transition of the cluster, creates a
population depletion in the ground state. The in-
frared absorption is then detected as a dip in the
fluorescence signal, because of the population de-
pletion. In our experiments, tunable IR was gen-
erated by a difference frequency generation (DFG)
with a LiNbO3 crystal, between the second har-
monic of a Nd:YAG laser (Quanta-Ray GCR/230)
and the output of the Nd:YAG pumped dye laser
(Continumm ND 6000) (DCMdye; wavelength
region 650–665 nm). The time delay between the
IR and UV lasers was 50 ns. The existence of
isomers, even in the case of binary clusters, has
been shown in many cases. To probe such an
eventuality, we employed IR–UV hole burning
spectroscopy. In this technique, the LIF excitation
spectrum for the region of interest is first recorded.
Following, the IR pulse tuned to a vibrational
transition of a particular isomer is introduced 50
ns prior to the UV laser pulse and the LIF exci-
tation spectrum is recorded again, under identical
conditions. In the presence of more than one iso-
mer, the LIF intensity diminishes for the transi-
tions corresponding to the isomer, to whose
vibrational transition the IR laser is tuned to and
remains unchanged for the rest of the isomers.
Subtracting the hole-burning spectrum from the
LIF excitation spectrum allows us to identify the
relevant transitions.
3. Results and discussion
DFT-based calculations were carried out with
Becke’s 3-parameter hybrid exchange correlation
functional (B3LYP) [27] using GAUSSIAN-98
[28] to determine the structure of all binary clus-
ters. The basis set used for this purpose was a split
valance 6-31++G(d, p), with diffuse and polariza-
tion functions on heavy atoms as well as hydro-
gens. It has been shown in several instances that
3.1. Clusters of phenol
The identification of the electronic transition
corresponding to the binary cluster is the first step
in the spectroscopic characterization, which was
achieved by recording the LIF excitation spec-
trum. Fig. 1(a) shows the S1 S0 LIF excitation

G.N. Patwari et al. / Chemical Physics 283 (2002) 193–207
197
bound complex. In both the cases, several new
transitions accompanying band origin were as-
signed as vibronic transitions on the basis of IR–
UV hole-burning spectroscopy presented later.
The shift in the electronic transition of phenol
upon cluster formation usually suggests the pos-
sible binding site, however, exceptions are known
[32]. As mentioned earlier, the shift in the hydride
stretching vibration of the proton donor is the best
spectroscopic tool for investigating hydrogen
bonding. Therefore, the IR spectra of bare phenol
and its clusters was recorded using FDIR spec-
troscopy. Fig. 2 shows the FDIR spectra of bare
phenol and its complexes with BTMA and
BDMA. The single transition at 3657 cmꢂ1 in Fig.
2(a) corresponds to the O–H stretching vibration
of phenol and is in excellent agreement with that
reported elsewhere [26,33]. The observed structure
is due to ambient water vapor absorption, which
modulates the IR power. The FDIR spectrum of
phenol–BTMA is shown in Fig. 2(b), which also
shows a single, slightly broad transition at
3514 cmꢂ1 for the O–H stretching vibration of the
complex, shifted by 143 cmꢂ1 to a lower frequency
from that of bare phenol. This value, once again, is
marginally larger than that of phenol–H2O (133
cmꢂ1) [31] and is in accord with observed red shift
in the electronic transition. Both, the red shift in
Fig. 1. LIF excitation spectrum of phenol in the presence of (a)
BTMA and (b) BDMA. Peaks marked ‘‘A’’ and ‘‘B’’ in (a) and
(b) are the respective band origin transitions of the binary
clusters. The transitions marked ‘‘P’’, ‘‘W’’, and ‘‘D’’ are due to
bare phenol, phenol–H2O, and phenol dimer, respectively.
Transitions indicated by an asterisk in (b) are the vibronic
bands of phenol–BDMA.
spectrum of phenol in the presence of BTMA.
Several new transitions appeared after the addition
of BTMA to the sample holder, of which, the
lowest energy transition marked ‘‘A’’ at 35,974
cmꢂ1 is the band origin of the phenol–BTMA
complex. The mass number of the ion responsible
for this signal was measured as 167 amu using a
quadrupole mass spectrometer with REMPI
scheme [22]. It thus corresponds to 1:1 cluster of
phenol and BTMA. The origin band of phenol–
BTMA is shifted by 384 cmꢂ1 to the red from the
corresponding band of bare phenol. The observed
red shift in the electronic transition is marginally
larger than that of phenol–H2O complex
[355 cmꢂ1; marked ‘‘W’’ in Fig. 1(a)] and clearly
suggests that phenol is acting as a proton donor to
BTMA, similar to phenol–H2O [31]. Fig. 1(b)
shows the LIF excitation spectrum of phenol in the
presence of BDMA. In this case, the peak marked
‘‘B’’ at 36,259 cmꢂ1 is the new lowest energy
transition, which accompanies six more transi-
tions, marked with asterisk. The red shift in the
band origin transition for phenol–BDMA is only
89 cmꢂ1, and suggests a possibility of p-electron
Fig. 2. FDIR spectra of (a) bare phenol, (b) phenol–BTMA,
and (c) phenol–BDMA.

198
G.N. Patwari et al. / Chemical Physics 283 (2002) 193–207
the electronic transition and the shift in the O–H
stretching vibration to a lower frequency in the
case of phenol–BTMA are well correlated with
each other and affirm that phenolic O–H is hy-
drogen bonded to BTMA.
from the case of phenol–BTMA, these two spectral
shifts do not correlate well with each other. A
survey of various hydrogen bonded phenol clusters
points towards the fact that the electronic transi-
tion shifts to the red when phenol is a proton
donor, while it shifts to blue when phenol acts as a
proton acceptor. On the other hand, only clusters
in which phenol acts as a proton donor show a
substantial shift in the O–H stretching to a lower
Fig. 2(c) shows the FDIR spectrum of phenol–
BDMA, in which the peak at 3272 cmꢂ1 is
assigned to the N–H stretching of the BDMA
moiety, on the basis of energy considerations.
Since the gas phase IR spectrum of BDMA in not
known, it was not possible to compare it with the
N–H stretching of free BDMA. An intense tran-
sition at 3483 cmꢂ1 was assigned to the phenolic
O–H stretching of the cluster, which is shifted by
174 cmꢂ1 to a lower frequency. The shift in the O–
H stretching vibrations is comparable to several
other hydrogen bonded complexes of phenol [34]
and implies that phenolic O–H is hydrogen bon-
ded to BDMA. Furthermore, four bands observed
on the higher frequency side of the O–H stretching
vibration (19, 31, 44, and 47 cmꢂ1) were assigned
as combination bands of intermolecular vibrations
on the O–H stretching vibration.
frequency. In the case of phenol–ðH2OÞ cluster,
2
which forms a cyclic structure in which phenol acts
as both a proton donor and an acceptor, the red
shift in the electronic transition is only 133 cmꢂ1
while the O–H stretching frequency shows a large
shift of 269 cmꢂ1 to a lower frequency [31]. The
trend in the present phenol–BDMA is very similar
to that of phenol–ðH2OÞ cluster. Extending the
2
analogy it is reasonable to assume that phenol–
BDMA forms a cyclic structure.
As can be seen from Fig. 1, several new tran-
sitions appear in the LIF excitation spectra of
phenol–BTMA and phenol–BDMA. In order to
check for the possibility of various isomers in each
spectrum, we carried out IR–UV hole-burning
spectroscopy, and the results are presented in Fig.
3. Trace I of Fig. 3(a) shows the LIF excitation
spectrum of phenol–BTMA on an expanded scale,
In the case of phenol–BDMA, the red shift of
89 cmꢂ1 in the electronic transition is rather small,
while the O–H stretching frequency shows a large
shift of 174 cmꢂ1 to a lower frequency. Different
Fig. 3. IR–UV hole-burning spectra (a) phenol–BTMA and (b) phenol–BDMA. In each case trace I is the LIF excitation spectrum
with IR off, trace II is the LIF excitation spectrum recorded by pumping the O–H stretching vibration prior to UV excitation (IR–UV
hole-burning spectrum) and trace III is the difference spectrum obtained by subtracting trace II from trace I.

G.N. Patwari et al. / Chemical Physics 283 (2002) 193–207
199
and trace II the IR–UV hole-burning spectrum,
which was recorded by pumping the O–H
The scaling factor of 0.955 was chosen to match
the experimental O–H stretching frequency at
3657 cmꢂ1, and was used throughout. The stabi-
lization energies of various clusters are listed in
Table 1. In the case of phenol–BTMA, we found
only one minimum on the potential energy surface
(PES), irrespective of the starting geometry for the
calculation, which is shown in Fig. 4(a) [complex
A]. This complex has the stabilization energy of
4.44 kcal/mol, in which the phenolic O–H group
interacts with the B–H group of BTMA forming a
pair of bifurcated dihydrogen bonds of the type
O–Hꢀ ꢀ ꢀðH–BÞ2. In these two dihydrogen bonds the
stretching vibration of the complex at 3514 cmꢂ1
,
prior to the UV excitation. It is evident from the
trace II that apart from the band origin all the
transitions have diminished intensities, except
those already assigned to the other clusters and
bare phenol. This clearly demonstrates the pres-
ence of only one isomer of phenol–BTMA. Trace
III is the resultant of subtracting trace II from
trace I, which is the LIF excitation spectrum of
phenol–BTMA, exclusively, and shows a rich
vibronic activity. A similar exercise was carried out
on phenol–BDMA, shown in Fig. 3(b). It also
indicates that all six new transitions belong to only
one isomer. Further, couple of extra peaks, which
can be seen in traces I and II are not due to phe-
nol–BDMA, as confirmed by mass resolved exci-
tation spectrum [15]. Trace III of Fig. 3(b) is the
LIF excitation spectrum of pure phenol–BDMA,
which also shows a considerable vibronic activity
in the low energy region.
ꢀ
ꢀ
Hꢀ ꢀ ꢀH contact distances are 1.87 A and 2.25 A,
while the corresponding B–Hꢀ ꢀ ꢀH angles of 103.0°
and 84.1° and O–Hꢀ ꢀ ꢀH angles being 151.4° and
143.0°, respectively. Clearly, these geometrical
parameters are in accord with those observed in
crystals of borane-amines reported in CSD [10]
and reiterate the formation of a dihydrogen bon-
ded complex.
In the case of phenol–BDMA, we found three
minima on the PES whose structures are displayed
in Figs. 4(b)–(d) with the corresponding stabiliza-
tion energies being 6.22, 4.37, and 3.58 kcal/mol,
respectively. From the energetics it is easy to re-
alize that only the most stable isomer will be sig-
nificantly populated in supersonic jets with the
nozzle temperature of 315 K. This is very much in
agreement with the experimental observation of
only one phenol–BDMA complex. Fig. 4(b) shows
the structure of the most stable phenol–BDMA
complex [B], which forms a six-member ring of O–
H–H–B–N–H atoms, with phenol acting both as a
proton donor and an acceptor. Even in this case
From the experimental results presented above,
it is very clear that phenol is hydrogen bonded to
both BTMA and BDMA. If one examines the
structures of BTMA and BDMA, one could in-
tuitively infer that phenol is dihydrogen bonded to
BTMA and BDMA, since B–H group is the only
sterically accessible proton acceptor in these two
molecules. To verify this and to determine the in-
termolecular structure we carried out DFT calcu-
lations. For the phenol monomer (figure not
shown), the calculated structure has the O–H
group in the plane of the benzene ring and the
unscaled O–H stretching frequency is 3829 cmꢂ1
.
Table 1
Stabilization (ZPE and BSSE corrected) energies of phenol complexes with BTMA and BDMA; also listed are the calculated
(observed) vibrational frequencies
DEa (kcal molꢂ1
)
DEb (kcal molꢂ1
)
mOH (cmꢂ1
)
mNH (cmꢂ1
)
Phenol
BDMA
–
–
3657 (3657)
–
–
–
–
3333
–
Phenol–BTMA (A)
Phenol–BDMA (B)
Phenol–BDMA (C)
Phenol–BDMA (D)
4.44
6.22
4.37
3.58
4.61
6.42
4.54
3.83
3505 (3514)
3480 (3483)
3506
3270 (3271)
3330
3277
3653
a With 100% BSSE correction.
b With 50% BSSE correction.

200
G.N. Patwari et al. / Chemical Physics 283 (2002) 193–207
(a)
(b)
(c)
(d)
Fig. 4. Calculated structures of (a) phenol–BTMA and (b)–(d) phenol–BDMA. Distances are given in Angstroms, and labels B, N, and
O correspond to boron, nitrogen and oxygen atoms, respectively. See text and Table 1 for details.
the O–H group of phenol is interacting with the
B–H group of BDMA forming a O–Hꢀ ꢀ ꢀH–B
dihydrogen bond, with the Hꢀ ꢀ ꢀH contact distance
complex C] has an intermolecular structure almost
identical to the phenol–BTMA with the two
ꢀ
Hꢀ ꢀ ꢀH contact distances of 1.86, 2.30 A, B–Hꢀ ꢀ ꢀH
ꢀ
of 1.87 A, B–Hꢀ ꢀ ꢀH angle of 100.0° and the
angles of 100.0°, 82.9° and the O–Hꢀ ꢀ ꢀH angles
being 150.1°, 142.9°. The third [Fig. 4(d); complex
D] is a conventional hydrogen bonded complex,
where the phenol acts as an acceptor to N–H of
BDMA, with the hydrogen bonding distance of
O–Hꢀ ꢀ ꢀH angle being 143.4°. Additionally, a
conventional hydrogen bond is also formed
between the N–H of BDMA moiety and the
phenolic oxygen. For this hydrogen bond, phenol
acts a proton acceptor, with the hydrogen bonded
ꢀ
2.07 A and N–Hꢀ ꢀ ꢀO angle of about 176.5°.
ꢀ
distance of 2.15A and N–Hꢀ ꢀ ꢀH angle being
As mentioned earlier, the agreement between
the experimental and calculated vibrational fre-
quencies in the hydride stretching region serves as
the basis for the structural assignment. The com-
parison of experimental and calculated vibrational
139.2°. This structure is very much in accordance
with the one predicted earlier on the basis of
spectral shifts in the electronic and O–H stretching
transitions. The second (local) minimum [Fig. 4(c);

G.N. Patwari et al. / Chemical Physics 283 (2002) 193–207
201
3.2. Clusters of aniline
spectra is presented in Fig. 5. Table 1 also lists
calculated and observed vibrational frequencies
for various clusters. Trace I of Fig. 5(a) shows the
experimental spectrum of phenol–BTMA, which is
identical to one shown in Fig. 2(b), but is pre-
sented in an inverted ordinate, to reflect the IR
absorption. The corresponding calculated spec-
trum, given in trace II, shows the O–H stretching
vibration at 3505 cmꢂ1, an excellent agreement
with the experimental value of 3514 cmꢂ1. Fig.
5(b) shows the comparison for phenol–BDMA, in
which the calculated spectrum of the most stable
complex, B, is in excellent agreement with the ex-
perimental spectrum, both in terms of frequencies
and intensities. This clearly confirms that phenol–
BDMA indeed has a cyclic structure in which
phenol is both a (di)hydrogen bond donor and a
hydrogen bond acceptor.
Aniline is the simplest aromatic amine, and thus
is a good model system to investigate hydrogen
bonding involving N–H bonds. Spectroscopic in-
vestigation of hydrogen bonded clusters of aniline
via electronic excitation presents an unique chal-
lenge due to the drastic structural change upon
electronic excitation. Most of the solvents, such as
H2O, MeOH, which act as good proton acceptors
for phenol [34] reverse their roles as proton donors
in the case of aniline, since it is a much stronger
proton acceptor than a donor, unlike phenol.
Though there are several reports of phenol–H2O
using electronic spectroscopy, however, on the
contrary aniline–H2O or any other cluster, in
which aniline acts as a proton acceptor [35], has
not been identified in the gas phase using elec-
tronic spectroscopy, which is believed to be due to
unfavorable Franck–Condon factors. However,
on the other hand several binary complexes of
aniline, in which it acts as a proton donor, have
been reported [35–42]. Since there are no acidic
protons and B–H group is the only accessible
proton acceptor in BTMA, we expect BTMA to
form a dihydrogen bonded complex with ani-
line.
The LIF excitation spectrum of aniline in the
presence of BTMA in the energy range of 33,200–
34,100 cmꢂ1 is shown in Fig. 6. Picking up the
lowest energy transition corresponding to the an-
iline–BTMA binary cluster was not an easy task
due to a large spread in the Franck–Condon re-
gion, with a very long progression of 20 cmꢂ1. The
band origin transition was identified at 33,325
cmꢂ1 with the aid of IR–UV hole-burning spec-
troscopy presented later. The intensity profile
suggests that the progression should involve more
than one vibrational mode of the excited state and
also indicates a drastic change in the structure
upon electronic excitation, typical to aniline clus-
ters [35–38]. The band origin of aniline–BTMA is
shifted to the red by 705 cmꢂ1, and is thus in the
same league as hydrogen bonded aniline–
OðCH3Þ2 ð666 cmꢂ1Þ [38], aniline–NH3ð889 cmꢂ1
Þ
[35], and aniline–NðCH3Þ ð876 cmꢂ1Þ [36] and
3
signifies that aniline is indeed acting as a proton
donor to BTMA, as expected.
Fig. 5. Comparison of experimental and calculated IR spectra
of (a) phenol–BTMA and (b) phenol–BDMA.

202
G.N. Patwari et al. / Chemical Physics 283 (2002) 193–207
with that reported elsewhere [35]. Fig. 7(b) shows
the FDIR spectrum of aniline–BTMA in the same
frequency region. This spectrum was recorded by
probing the reasonably intense vibronic transition
100 cmꢂ1 blue to the band origin (marked ‘‘+’’ in
Fig. 6). Comparison with other hydrogen bonded
complexes of aniline suggests that the broad and
intense transition at 3387 cmꢂ1 is most likely due
to the hydrogen bonded N–H and the higher fre-
quency band at 3471 cmꢂ1 due to free N–H [33].
The shifts in the N–H stretching frequencies in the
case of aniline are not straightforward, as in the
case of O–H stretching frequency phenol. The
NH2 group of aniline has C2V local symmetry with
two coupled identical N–H bonds, which gives a
lower frequency symmetric ðmaÞ and a higher fre-
quency asymmetric ðmasÞ stretching vibrations.
Upon cluster formation the local symmetry is
lowered and the two N–H bonds are decoupled,
which would place the free N–H stretching at
3475 cmꢂ1 (median of ms and mas), a value very
close to that observed in the case of aniline–
BTMA (3471 cmꢂ1). Therefore the difference in
the free and hydrogen bonded stretching N–H
frequencies is the real shift in the N–H stretching
upon complex formation, which is 84 cmꢂ1 to a
lower frequency in the present aniline–BTMA.
Apart from the two dips corresponding to the
hydrogen bonded and the free N–H stretching of
aniline–BTMA, two fluorescence-enhanced signals
can also be seen in Fig. 7(b), which correspond to
the symmetric and asymmetric stretching of NH2
group of bare aniline. The enhanced signal is due
to the fluorescence from the higher vibronic levels
of aniline pumped by UV laser following IVR
from the IR excited N–H stretching vibrational
levels in the ground state [43].
Fig. 6. LIF excitation spectrum of aniline in the presence of
BTMA. Very week band origin transition is marked with an
asterisk, the transition marked with ‘‘+’’ was used for recording
the FDIR spectrum, and the one marked with ‘‘#’’ is due to
some impurity.
Fig. 7(a) shows the FDIR spectrum of bare
aniline in the N–H stretching region, in which the
two transitions at 3422 and 3509 cmꢂ1 correspond
to NH2 symmetric and asymmetric stretching vi-
brations, respectively, and are in good agreement
The results of IR–UV hole-burning spectro-
scopy are presented in Fig. 8. The upper panel
shows an expanded view of the LIF excitation
spectrum, followed by the hole-burning spectrum,
which was recorded by pumping the hydrogen
bonded N–H stretching vibration at 3387 cmꢂ1
.
The broad background in the above two spectra
may be due to higher clusters of aniline and
BTMA and/or higher clusters of aniline. The third
panel shows the difference spectrum, obtained by
subtracting second from the first, which gives the
Fig. 7. FDIR spectra of (a) aniline and (b) aniline–BTMA,
presented along with the calculated IR spectra. See text for
details.

G.N. Patwari et al. / Chemical Physics 283 (2002) 193–207
203
(a)
(a)
(b)
#
#
(c)= (a) - (b)
*
(b)
33300
33400
33500
33600
33700
33800
33900
Energy / cm-1
Fig. 8. (a) LIF excitation spectrum of aniline–BTMA with IR
off, (b) LIF excitation spectrum recorded by pumping the N–H
stretching vibration at 3387 cmꢂ1 prior to UV excitation (IR–
UV hole-burning spectrum), and (c) difference spectrum ob-
tained by subtracting trace II from trace I. Peak marked with
‘‘#’’ in (a) and (b) is due to some impurity, and can also be seen
in Fig. 6.
(c)
LIF excitation spectrum of aniline–BTMA, ex-
clusively. This spectrum distinctly shows a weak
band origin transition at 33,325 cmꢂ1 and also
confirms the presence of only one species of ani-
line–BTMA.
We also recorded the LIF excitation spectrum
of aniline in the presence of BDMA covering the
energy range of about 1000 cmꢂ1 to the red and
2000 cmꢂ1 to the blue of aniline band origin.
However, we could not identify any new transi-
tions indicating the formation of the aniline–
BDMA binary cluster. The calculated structures of
aniline–BTMA and aniline–BDMA are displayed
in Fig. 9 and their stabilization energies are listed
in Table 2. Once again we found only one mini-
mum on the PES of aniline–BTMA. Though we
started our calculations starting from several ini-
tial geometries generated randomly, all those
converged gave the same minimum, which is
shown in Fig. 9(a) [E]. In this structure one of the
N–H’s of aniline is interacting with one B–H
group of BTMA, forming N–Hꢀ ꢀ ꢀH–B dihydrogen
Fig. 9. Calculated structures of (a) aniline–BTMA and (b), (c)
aniline–BDMA complexes. Distances are given in Angstroms,
and labels B and N correspond to boron and nitrogen atoms,
respectively. See text and Table 2 for details.
148.8°, very similar to borane-amines [10]. Addi-
tionally, a week hydrogen bond of length 2.67 A is
ꢀ
formed between the nitrogen lone pair of aniline
and a methyl proton of BTMA. The stabilization
energy of this complex is 2.39 kcal/mol, about 2
kcal/mol smaller than its phenol counterpart,
which can be attributed to the lower acidity of
N–H of aniline, than that of the O–H of phenol. In
the case of aniline–BDMA we found two minima,
ꢀ
bond with a Hꢀ ꢀ ꢀH contact distance of 2.06 A,
B–Hꢀ ꢀ ꢀH angle of 131.6° and N–Hꢀ ꢀ ꢀH angle of

204
G.N. Patwari et al. / Chemical Physics 283 (2002) 193–207
Table 2
Stabilization (ZPE and BSSE corrected) energies of aniline complexes with BTMA and BDMA; also listed are the calculated (ob-
served) vibrational frequenciesa
DEb
DEc
m1 (cmꢂ1
)
m2 (cmꢂ1
)
mNH (cmꢂ1
)
(kcal molꢂ1
)
(kcal molꢂ1
)
Aniline
BDMA
–
–
3414 (3422)
–
3516 (3509)
–
–
–
–
3333
–
Aniline–BTMA (E)
Aniline–BDMA (F)
Aniline–BDMA (G)
2.39
2.46
5.55
2.57
2.60
5.82
3381 (3387)
3378
3367
3494 (3471)
3490
3467
3331
3162
a m1 and m2 are the symmetric and asymmetric stretching vibrations of NH2 group in the case of bare aniline and lower and higher
frequency vibrations for the clusters.
b With 100% BSSE correction.
c With 50% BSSE correction.
which are also shown in Fig. 9. The first one [Fig.
9(b); F] has the structure and stabilization energy
very similar to aniline–BTMA, and does not
warrant further discussion. The second is a con-
ventional hydrogen bonded complex [Fig. 9(c); G]
with N–H of BTMA moiety interacting with the
nitrogen of aniline. In this case the hydrogen
electronic excitation, corresponding to vibrations
involving intermolecular motions. This feature has
not been distinctly noticeable in the case of hy-
drogen bonded complexes of phenol. Spectro-
scopic identification of phenol–BTMA, phenol–
BDMA and aniline–BTMA dihydrogen bonded
clusters, in the gas phase, was achieved by exam-
ining the spectral shifts in the electronic and hy-
dride (O–H and N–H) stretching transitions
relative to the bare molecules. These are the same
parameters used for characterization of conven-
tional hydrogen bonded complexes. Therefore a
comparison between the dihydrogen bonding and
the conventional hydrogen bonding is in order.
Table 3 lists the shifts in the electronic and O–H
stretching transitions of phenol–BTMA and phe-
nol–BDMA along with various other hydrogen
bonded clusters of phenol. In the case of phenol
complexes with various proton acceptors, such as
H2O, MeOH, NH3, amines, and other proton
ꢀ
bonded Nꢀ ꢀ ꢀH distance is 2.09 A and the
N–Hꢀ ꢀ ꢀH angle is bent with a value of 154.6°. The
stabilization energy of this complex is 5.55 kcal/
mol, about 3 kcal/mol larger than the first one.
This would imply that in our experiments only the
most stable aniline–BDMA [G] would be popu-
lated exclusively, which could not be observed like
many other aniline clusters, in which aniline is a
proton acceptor. Fig. 7 also shows the comparison
between experimental and calculated spectra for
aniline and aniline–BTMA. Further, Table 2 lists
N–H stretching frequencies of aniline, aniline–
BTMA, and aniline–BDMA. From Fig. 7 and
Table 2 it is very clear that the agreement between
the experimental and calculated spectra for both
aniline and aniline–BTMA is reasonable. This es-
tablishes the formation of N–Hꢀ ꢀ ꢀH–B dihydrogen
bonded complex between aniline and BTMA.
Table 3
Observed spectral shifts in the electronic transition and O–H
stretching vibration of various phenol clusters with respect to
bare phenol
Dmel
(cmꢂ1
DmOH
(cmꢂ1
)
Type
)
3.3. Comparison with conventional hydrogen bonded
clusters
Phenol–BTMAa
Phenol–H2Ob
)384
)355
)415
)631
)89
)143
)133
)201
)363
)174
)269
Linear
Linear
Linear
Linear
Cyclic
Cyclic
Phenol–MeOHb
One of the most striking features observed in
the LIF excitation spectra of phenol–BTMA,
phenol–BDMA and other dihydrogen bonded
complexes [17] is that they exhibit a rich vibronic
activity in the low frequency region upon the
b
Phenol–NH3
Phenol–BDMAa
b
Phenol–ðH2OÞ
)133
2
a This work.
b Ref. [34].

G.N. Patwari et al. / Chemical Physics 283 (2002) 193–207
205
acceptors, a near linear correlation exists between
the proton affinity of the solvent and the red shift
in the electronic transition and the shift in the O–H
stretching vibration of phenol moiety to a lower
frequency [24,34]. In the case of phenol–BTMA,
both the red shift in electronic transition and low
frequency shift in the O–H stretching vibrations
are higher than that of phenol–H2O and lower
than phenol–MeOH. On the basis of spectral
shifts, the proton affinity of the dihydrogen
bonding site of BTMA should lie between the 165
kcal/mol (H2O) and 180 kcal/mol (MeOH), and
can be roughly estimated as ꢃ168 kcal/mol. The
low frequency shift of the phenolic O–H stretching
vibration in phenol–BDMA is 174 cmꢂ1, which is
about 30 cmꢂ1 higher than phenol–BTMA. Since
the shift in the O–H vibrational frequency is al-
most due to the proton affinity of the acceptor, the
proton affinity of dihydrogen bonding site in this
case can be estimated as ꢃ174 kcal/mol. It appears
that the difference in the proton affinity of BTMA
and BDMA is probably due to the electronic
properties of trimethyl and dimethyl groups.
However, if one examines the geometries and sta-
bilization energies of the purely dihydrogen bon-
ded complexes of phenol with BTMA and BDMA
[A and C], it is easy to realize that a difference of 6
kcal/mol in proton affinity cannot be explained by
substituting a methyl group with a proton. In the
case of higher water complexes of phenol, the shift
of phenolic O–H stretching to a lower frequency
are: phenol–H2O, phenol–ðH2OÞ , phenol–ðH2OÞ ,
tially means that the cooperative effect is largest
when phenol is both proton donor and acceptor.
Extending the analogy it can be concluded that the
larger shift in the case of phenol–BDMA, relative
to phenol–BTMA, is due to the cooperative effect
induced by the formation of a cyclic complex.
We also calculated the structures and vibra-
tional frequencies of convential hydrogen bonded
phenol–H2O and phenol–MeOH using the same
level of theory [B3LYP/6-31++G(d, p)] as the di-
hydrogen bonded complexes. The calculated sta-
bilization energy for phenol–H2O is 4.60 kcal/mol,
a value marginally higher than phenol–BTMA,
contrary to the observed spectral shifts. Further,
the calculated phenolic O–H stretching frequency
of phenol–H2O ð3469 cmꢂ1Þ and phenol–MeOH
ð3415 cmꢂ1Þ are also much lower than the corre-
sponding experimental values (3524 and
3456 cmꢂ1). The same calculations predict the
symmetric, antisymmetric O–H stretching of H2O
in phenol–H2O and free O–H stretching of MeOH
in phenol–MeOH to be 3635, 3745 and 3665 cmꢂ1
,
respectively, which are in excellent agreement with
the corresponding experimental values of 3650,
3748 and 3676 cmꢂ1. This indicates a systematic
over estimation of conventional hydrogen bonding
interaction energy relative to dihydrogen bonding
at present level of theory. Nevertheless, a linear
correlation between the hydrogen bond strength
and O–H frequency shift exists in all the complexes.
Unlike phenol, the number of hydrogen bonded
complexes of aniline reported are much sparse. As
noted earlier, in all the reported complexes aniline
acts as a proton donor, and can be classified into
two categories. The first in which the proton in-
teracts with a p cloud of aromatic molecules
forming an N–Hꢀ ꢀ ꢀp-type hydrogen bond, and the
second one is in the interaction with a lone pair of
electrons forming an N–Hꢀ ꢀ ꢀr-type hydrogen
bond. Table 4 lists the shifts in the electronic and
N–H vibrational transitions for various clusters of
aniline along with the present aniline–BTMA. The
shifts in the N–H vibrational transitions in Table 4
are defined as follows: DmS ¼ ms ꢂ m1 and
DmA ¼ mas ꢂ m2; where ms; mas are the symmetric,
asymmetric stretching of bare aniline and m1; m2 are
the two N–H stretches in the complex with m1 < m2.
The shifts in the electronic and vibrational transi-
2
3
and phenol–ðH2OÞ are 133, 269, 421, and
4
490 cmꢂ1, respectively [31]. This implies a drastic
increase in proton affinity of water with increase in
cluster size, which is due to the cooperative effect
of the hydrogen bonding network in these clusters.
Watanabe and Iwata [44] reported the geometries
and vibrational frequencies of several phenol–
ðH2OÞn clusters. For n P 2 the most stable struc-
ture is always the one in which the hydrogen
bonding network wraps around phenolic O–H
group forming a cyclic structure, with phenol
acting as both proton donor and acceptor. Com-
parison of O–H stretching frequencies reveals that
hydrogen bonding network does indeed increase
the shift towards lower frequency, with the maxi-
mum being in the most stable isomer. This essen-

206
G.N. Patwari et al. / Chemical Physics 283 (2002) 193–207
Table 4
Observed spectral shifts in the electronic transition and N–H stretching vibrations of various aniline clusters with respect to bare
aniline
Dmel ( cmꢂ1
)
DmS (cmꢂ1
)
DmA (cmꢂ1
)
Type
a
a
Aniline–BTMAb
)705
)666
)889
)876
)791
)277
)672
)436
)345
35
40
68
94
52
12
28
29
29
38
23
29
42
29
11
42
41
42
r
r
r
r
r
r
p
c
Aniline–OðCH3Þ2
c
Aniline–NH3
Aniline–NðC2H5Þ
c
3
Aniline–THFd
Aniline–furanc
Aniline–anilinec
Aniline–benzenec
Aniline–pyrrolec
p
p
a DmS ¼ ms ꢂ m1 and DmA ¼ mas ꢂ m2; m1 < m2 (see text for details).
b This work.
c Ref. [39].
d Ref. [37].
tions of aniline do not have a clear linear relation
ship between the proton affinity as in the case of
phenol. However, from Table 4 it is very clear that
the r-type complexes have much larger red shifts in
the electronic transition and DmS relative to the p-
type complexes, with an exception of aniline–furan
[38]. Moreover, it can also be seen from Table 4
that for all the three p-type complexes listed, DmS
and DmA are constant at 29 and 42 cmꢂ1, respec-
tively. Additionally, from the IR spectra of various
reported aniline clusters [35–42], following can be
noted: clusters with a r-type hydrogen bond show
a much broader and intense transition corre-
sponding to the hydrogen bonded N–H stretching
(m1) and a relatively week and narrow transition for
the free N–H stretching vibration (m2) [35–37]. On
the other hand, the p-type complexes show a
spectrum with two narrow transitions, very similar
to bare aniline [39–42]. The red shifts in the elec-
tronic transition (705 cmꢂ1) and the structure of
the IR spectrum (Fig. 7(b)) of aniline–BTMA are
comparable to other r-type hydrogen bonded
clusters of aniline, this complex can be classified as
r-type, similar to phenol–BTMA.
phenol/aniline was presented. The spectral shifts,
both in the electronic as well as vibrational tran-
sitions, along with the DFT calculations, allowed
the structural assignment. The excellent agreement
between the observed and calculated vibrational
spectra indicated that B3LYP/6-31++G(d, p) level
of calculation is quite reliable in describing dihy-
drogen bonded systems. From the comparison
between the present dihydrogen bonded complexes
with other hydrogen bonded clusters of phenol
and aniline following inferences can be made: (a)
The spectral characteristics and properties of di-
hydrogen bonded systems are very similar to
conventional (r) hydrogen bonding. (b) Though
the dihydrogen bonding is very different structur-
ally, especially the strongly bent angles as opposed
to almost linear in the case of conventional hy-
drogen bonding, both interactions can be modeled
within the same theoretical framework. (c) Finally,
the comparison allowed evaluating the gas phase
basicity of borane-amines, which have proton af-
finity similar to H2O and MeOH.
In this paper we have investigated the hydride
stretching vibrations of proton donors to the di-
hydrogen bonds. For the comprehensive under-
standing of dihydrogen bonding, the investigation
of B–H stretching vibrations in borane-amines is
very essential. Unfortunately, the current IR
source in our laboratory does not cover the B–H
4. Conclusions
The formation of dihydrogen bonded com-
plexes in the gas phase between borane-amines and
stretching frequency region of 2300–2400 cmꢂ1
.

G.N. Patwari et al. / Chemical Physics 283 (2002) 193–207
207
[18] G.N. Patwari, T. Ebata, N. Mikami, J. Chem. Phys. 116
(2002).
Efforts are being made in this direction and
hopefully we will be able to investigate B–H
stretching in the near future.
[19] Q. Liu, R. Hoffmann, J. Am. Chem. Soc. 117 (1995) 10108.
[20] J.L. Atwood, G.A. Koutsantonis, F. Lee, C.L. Raston,
J. Chem. Soc., Chem. Commun. (1994) 91.
[21] S.A. Kulkarni, J. Phys. Chem. A 103 (1999) 9330.
[22] G.N. Patwari, T. Ebata, N. Mikami, J. Phys. Chem. A 105
(2001) 10753.
Acknowledgements
[23] J.P. Campbell, J.W. Hwang, V.G. Young Jr., R.B. Von
Dreele, C.J. Cramer, W.L. Gladfelter, J. Am. Chem. Soc.
120 (1998) 521.
G.N.P. thanks J.S.P.S. for the fellowship. T.E.
acknowledges the support in part by the Grant-in-
Aids for Scientific Research (No. 12640483) by
J.S.P.S. N.M. acknowledges the support by Re-
search for the Future, Photoscience (JSPS-RFTS-
98P-01203). Authors also wish to thank Drs. A.
Fujii, H. Ishikawa, and T. Maeyama for helpful
suggestions and discussion.
[24] In the case of hydrogen bonded complexes of phenol with
various proton acceptors, the shift in the O–H stretching
vibration of phenol to a lower frequency, depends only on
the strength of interaction and not the type of interaction.
A. Fujii, T. Ebata, and N. Mikami, submitted.
[25] G.C. Pimentel, A.L. McClellan, The Hydrogen Bond,
Freeman, San Francisco, 1960.
[26] S. Tanabe, T. Ebata, M. Fujii, N. Mikami, Chem. Phys.
Lett. 215 (1993) 347.
[27] A.D. Becke, Phys. Rev. A 38 (1988) 3098;
C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.
[28] M.J. Frisch et al., GAUSSIAN 98, Revision A.7, Gauss-
ian, Inc., Pittsburgh, PA, 1998.
References
[1] G.A. Jeffery, W. Saenger, Hydrogen Bonding in Biological
Structures, Springer, Berlin, 1994.
[29] C.J. Cramer, W.L. Gladfelter, Inorg. Chem. 36 (1997)
5358.
[2] J. Wanna, J.A. Menapace, E.R. Bernstein, J. Chem. Phys.
85 (1986) 1795.
[30] K.S. Kim, P. Tarakeshwar, J.Y. Lee, Chem. Rev. 100
(2000) 4145.
[3] T.S. Zwier, Ann. Rev. Phys. Chem. 47 (1996) 205.
[4] R.C. Stevans, R. Bau, R. Milstein, O. Blum, T.F. Koetzle,
J. Chem. Soc., Dalton Trans. (1990) 1429.
[5] L.S. Vander Sluys, J. Eckert, O. Eisenstein, J.H. Hall, J.C.
Huffman, S.A. Jackson, J.F. Koetzle, G.J. Kubas, P.J.
Vergamini, K.G. Caulton, J. Am. Chem. Soc. 112 (1990)
4381.
[31] T. Watanabe, T. Ebata, S. Tanabe, N. Mikami, J. Chem.
Phys. 105 (1996) 408.
[32] The shift in the electronic transition in the case of cyclic
complexes of phenol are rather low and are usually
misleading in terms of interaction strength. For example
see Ref. [31].
[6] G. Olah, K. Wade, R.E. Williams, Electron Deficient
Boron and Carbon Clusters, Wiley, New York, 1991.
[7] E.L. Lippert, W.N. Lipscomb, J. Am. Chem. Soc. 78
(1956) 503.
[33] H.D. Bist, J.C.D. Brand, D.R. Williams, J. Mol. Spectrosc.
24 (1967) 413.
[34] A. Iwasaki, A. Fujii, T. Watanabe, T. Ebata, N. Mikami,
J. Phys. Chem. 100 (1996) 16053.
[8] W.T. Klooster, T.F. Koetzle, P.E.M. Siegbahn, T.B.
Richardson, R.H. Crabtree, J. Am. Chem. Soc. 121
(1999) 6337.
[35] T. Nakanaga, K. Sugawara, K. Kawamata, F. Ito, Chem.
Phys. Lett. 267 (1997) 491.
[36] P.K. Chowdhury, K. Sugawara, T. Nakanaga, H. Takeo,
J. Mol. Struct. 447 (1998) 7.
[9] T.B. Richardson, S. deGala, R.H. Crabtree, P.E.M.
Siegbahn, J. Am. Chem. Soc. 117 (1995) 12875.
[10] R.H. Crabtree, P.E.M. Siegbahn, O. Eisenstein, A.L.
Rheingold, T.F. Koetzle, Acc. Chem. Res. 29 (1996) 348.
[11] S.A. Kulkarni, J. Phys. Chem. A 102 (1998) 7704.
[12] S.A. Kulkarni, A.K. Srivastava, J. Phys. Chem. A 103
(1999) 2836.
[37] P.K. Chowdhury, Chem. Phys. Lett. 319 (2000) 501.
[38] T. Nakanaga, N.K. Piracha, F. Ito, J. Phys. Chem. A 105
(2001) 4211.
[39] T. Nakanaga, F. Ito, J. Phys. Chem. A 103 (1999)
5440.
[40] K. Sugawara, J. Miyawaki, T. Nakanaga, H. Takeo, G.
Lembach, S. Djafari, H.D. Barth, B. Brutschy, J. Phys.
Chem. 100 (1996) 17145.
[13] S.J. Grabowski, Chem. Phys. Lett. 327 (2000) 203.
[14] P.L.A. Popelier, J. Phys. Chem. A 102 (1998) 1873.
[15] G.N. Patwari, T. Ebata, N. Mikami, J. Chem. Phys. 113
(2000) 9885.
[41] P.K. Chowdhury, K. Sugawara, T. Nakanaga, H. Takeo,
Chem. Phys. Lett. 285 (1998) 77.
[16] G.N. Patwari, T. Ebata, N. Mikami, J. Chem. Phys. 114
(2001) 8877.
[42] K. Kawamata, P.K. Chowdhury, F. Ito, K. Sugawara,
T. Nakanaga, J. Chem. Phys. A 102 (1998) 4788.
[43] G.N. Patwari, T. Ebata, and N. Mikami, submitted.
[44] H. Watanabe, S. Iwata, J. Chem. Phys. 105 (1996) 420.
[17] G.N. Patwari, T. Ebata, N. Mikami, J. Phys. Phys. A 105
(2001) 8462.