HCl and HCl-base adducts in silicalite channels as models of acid-base interactions in zeolites: An IR and theoretical study

Article abstract of DOI:10.1021/jp9816871

The IR spectroscopy of HCl adsorbed on silica and on silicalite and of HCl-B (B = (CH₃)₂O, (CH₃CH₂)₂O, THF) adducts formed in silicalite channels at a??210 K is described in detail. The whole subject is divided in three parts. The first part concerns the IR study of the HCl interaction with silica and silicalite. The formation of weak H-bonded adducts between silanols and HCl molecules has been observed; the number of HCl molecules in the adducts increases with the HCl dosage. The HCl molecules adsorbed on silicalite behave as partially hindered rotators. The experimental results are supported by ab initio and molecular dynamics calculations. In the second part, the IR study of HCl-B complexes (B = (CH₃)₂O, (CH₃CH₂)₂O, THF, and benzene) either adsorbed on silica or entrapped in silicalite channels is described. In the third part, the results obtained for HCl-B adducts in silicalite are compared with those obtained in cryogenic matrixes and in solution. It is inferred that the acid strength of HCI is appreciably influenced by the presence of the siliceous framework. This result allows discussion of the effect of long-distance interactions (framework effect) on the acid strength of structural Brnsted sites of zeolites. ? 1998 American Chemical Society References and Notes.

Full text of DOI:10.1021/jp9816871

J. Phys. Chem. B 1998, 102, 10753-10764
10753
HCl and HCl-Base Adducts in Silicalite Channels as Models of Acid-Base Interactions in
Zeolites: An IR and Theoretical Study
C. Paz e´ , B. Civalleri, S. Bordiga, and A. Zecchina*
Dipartimento di Chimica Inorganica, Chimica Fisica e Chimica dei Materiali, UniVersit a` di Torino,
Via P. Giuria 7, I-10125, Torino, Italy
ReceiVed: March 31, 1998
3 2 3 2 2
The IR spectroscopy of HCl adsorbed on silica and on silicalite and of HCl-B (B ) (CH ) O, (CH CH ) O,
THF) adducts formed in silicalite channels at ∼210 K is described in detail. The whole subject is divided in
three parts. The first part concerns the IR study of the HCl interaction with silica and silicalite. The formation
of weak H-bonded adducts between silanols and HCl molecules has been observed; the number of HCl
molecules in the adducts increases with the HCl dosage. The HCl molecules adsorbed on silicalite behave as
partially hindered rotators. The experimental results are supported by ab initio and molecular dynamics
3 2 3 2 2
calculations. In the second part, the IR study of HCl-B complexes (B ) (CH ) O, (CH CH ) O, THF, and
benzene) either adsorbed on silica or entrapped in silicalite channels is described. In the third part, the results
obtained for HCl-B adducts in silicalite are compared with those obtained in cryogenic matrixes and in
solution. It is inferred that the acid strength of HCl is appreciably influenced by the presence of the siliceous
framework. This result allows discussion of the effect of long-distance interactions (framework effect) on the
acid strength of structural Brønsted sites of zeolites.
1. Introduction
CHART 1
Protonic zeolites are catalysts active in a wide variety of acid-
catalyzed reactions. The acid strength of these systems can vary
over a wide range and can be assumed to depend on several
factors. These can be broadly divided into two (not fully
independent) groups: those primarily determined by the local
structure of the Brønsted sites and those influenced by the three-
dimensional structure of the zeolite. Among the first group, we
of the δ and γ bending modes). Above all, the ∆ νj (AH) shift is
particularly useful because it is proportional to the formation
mention the bridged structure of the OH Brønsted group, the
7
,8
enthalpy of the complex between AH and B. A very much
1
(
SiO) and (AlO) distances, and the angle R (Chart 1).
9
used probe molecule is CO, because of its small size and simple
Long distance interactions (such as those associated with
spectrum and because it allows discrimination between acid sites
of different strength. Many other probe molecules can also be
electrostatic interactions with the second, third, etc. shell atoms
of the framework and with the extraframework ions) are the
main contributors to the second group. These collective effects
9
used, either binuclear (N2, H2, NO) and polynuclear (ethers,
10
alcohols). Molecules such as NH3 and pyridine are also widely
are analogous to those which, in the liquid phase, are named
1
4,15
used
because they allow estimation of the amount of
2
“
solvent effects”. The evaluation of the total acid strength and
+
+
Brønsted sites, through formation of NH4 and PyH ions. On
the basis of a systematic study of the shifts induced by bases
with proton affinities in the range 118-204 kcal/mol, it has
of the relative role of the two groups of factors in determining
the acidity of the acid sites of the zeolites is certainly a problem
of primary importance.
1
6
been shown that it is possible to estimate the acid strength of
a Brønsted group in a way that is independent of the particular
probe molecule used to test the acidity. This method is the
extension to heterogeneous conditions of the Bellamy-Hallam-
A well-accepted method3 to study the acid strength of
Brønsted sites in protonic zeolites is based on the analysis, by
means of IR spectroscopy, of the perturbations that the interac-
tion with basic molecules B induces on the vibrational properties
of Brønsted sites, AH, and of the base B, due to the formation
of AH-B adducts. The spectroscopic variations that give the
most important information on the hydrogen-bonding interaction
are those relative to (i) the frequency changes of the internal
modes of the probe molecules (for instance, for a diatomic
molecule AB, the variation ∆ νj of the stretching mode ν(A-B)
and (ii) the modification of the frequencies of the AH modes
-6
1
1-13
Williams plot.
A comparison of the data relative to the strong acid sites of
different zeolites, all containing the common bridged Si(OH)Al
1
6
structure, has shown the following acidity trend:
H-Y (Si/Al ) 3) < H-MORD (Si/Al ) 5) =
H-â (Si/Al ) 12.5) = H-ZSM-5 (Si/Al ) 14)
(
the shift ∆ νj (AH) of the ν(AH) stretching mode and the shifts
The acidity trend illustrated is difficult to interpret directly,
given the changes in both framework type and silicon-to-
aluminum ratio. It is evident that a simple relation between the
acid strength and the local structure on one side and between
*
Author to whom correspondence should be addressed. Phone:
+
3911 6707537. Fax: +3911 6707855. E-mail: ZECCHINA@
SILVER.CH.UNITO.IT.
1
0.1021/jp9816871 CCC: $15.00 © 1998 American Chemical Society
Published on Web 12/03/1998

10754 J. Phys. Chem. B, Vol. 102, No. 52, 1998
Paz e´ et al.
the acid strength and the chemical composition of the zeolite
allowed both high-temperature treatments in vacuo and low-
temperature IR measurements to be performed in situ. The IR
measurements were performed at 2 cm resolution on a Bruker
66 instrument equipped with a MCT cryogenic detector.
(primarily determining the collective effects) on the other side
-1
is not clearly emerging. In fact, zeolites characterized by
isostructural Brønsted groups and roughly the same Si/Al ratio,
like H-Y and H-MORD, show different acidities, while
H-MORD and H-ZSM-5, characterized by very different Si/
Al ratios, have comparable acid strengths. On the basis of the
initial considerations, these observations lead immediately to
the following question: is it possible to estimate the relative
influences on the acid strength of the effects associated with
the local structure and of the long-distance interactions? A
preliminary way to gain information on the approximate
magnitude of this effect is to plan an experiment where the acid
properties of a well-known molecule with Brønsted acid
character (HCl) adsorbed in the channels and cavities of a
neutral zeolitic framework are probed by studying the interaction
with suitable basic molecules B. The comparison of the
spectroscopic perturbations induced by the formation of B-HCl
adducts in the channels of the neutral zeolitic framework with
those of the same molecule in vacuo and in a cryogenic matrix
can in fact give information about the effect that the long-
distance interaction associated with the atoms of the neutral
framework (considered as a matrix encaging the adducts or,
more approximately, a solid and not uniform solvent) exerts
on the acid properties of the encapsulated molecule. The
feasibility of such an experiment has been already demon-
All the experiments were carried out at T ) 214 ( 5 K,
chosen so that the amount of adsorbed HCl is sufficiently large
to give strong IR spectra. As a result of unaivodable fluctuations
of the temperature between 209 and 219 K in our experimental
setup, it was not possible to make the measurements of the
isosteric heat of sorption. Gravimetric experiments with a
vacuum microbalance were not performed due to the corrosive
nature of HCl.
The coadsorption experiments of HCl with (CH3)2O,
(CH3CH2)2O, THF, and benzene have been conducted in the
following way. After activation, the silica and/or the silicalite
samples previously cooled at 214 ( 5 K, were contacted with
HCl (p = 266 Pa), and the spectra were recorded after the
attainment of equilibrium. Then (CH3)2O, (CH3CH2)2O, THF,
and benzene (from an external reservoir) were allowed to
gradually flow in and the IR spectra of the B-HCl complexes
gradually formed inside the silicalite channels were taken at
fixed time intervals, until the total consumption of the signal
of adsorbed HCl was achieved.
It is worth noting that during the experiment only the sample
holder was kept at 214 ( 5 K, while all the remaining parts of
the cell were at room temperature. This means that the gas phase
in equilibrium with the pellet can be prevalently considered to
be at ∼300 K. Under the experimental conditions of pressure
and temperature (∼300 K), the amount of B-HCl adducts in
the 1:1 gaseous phase is negligible and does not appreciably
contribute to the IR spectra.
1
7
strated for HCl-CH3CN complexes, which have been syn-
thesized in situ in the channels of silicalite (i.e., a totally siliceous
zeolite with MFI structure). The key point of this experiment
was represented by the fact that the silicalite structure showed
a good resistance to chemical attack by HCl, so allowing the
HCl-CH3CN interaction to be studied in its purest form. From
the vibrational properties of this complex, the magnitude of the
solvent effect exerted by the neutral MFI structure on the
complexes was then estimated. In this paper, we extend the
investigation to other basic molecules such as (CH3)2O,
2
.2. Calculation. Molecular dynamic calculations were
obtained using the classical mechanics Discover 95.0 program
21
from Molecular Simulations. The cff91 (version 3.0) force field
22
developed by Sauer and Hill, based on fit to ab initio energy
(CH3CH2)2O, THF, and benzene, and the whole body of results
surface data, was used for all the atoms. The extension of this
force field to the hydrogen and chlorine atoms is arbitrary. The
periodic structures were optimized by means of constant pressure
lattice energy minimization with respect to atomic Cartesian
coordinates and cell parameters. The system energy was
evaluated using a cutoff of 9.5 Å for van der Waals interactions
and the Ewald method for Coulombic interactions (to an
accuracy of 0.025 kcal/mol). Minimization was performed using
the steepest descent method for a few steps, followed by the
conjugate gradient method and then by the quasi-Newton-
Raphson method (BFGS), until the maximum derivative of the
energy with respect to the Cartesian coordinates was less than
is discussed. A detailed discussion of the state of HCl adsorbed
on silicalite probed with atomistic and ab initio simulation
techniques is also added. The reader must acknowledge that
this investigation has only a limited scope, as it cannot be
extended to real zeolitic systems, where the effect of the ionic
charge localized on the extraframework ions and on the zeolite
3+
framework, containing Al , is certainly playing an important
role.
2
. Experimental and Calculation Methods
.1. Experimental. The silicalite sample (kindly supplied by
2
0
.001 kcal/Å. The first minimization step (at variable cell
ENICHEM laboratories, Novara, Italy) was synthesized in a Na-,
K-, Al-, Fe-free form (<20 ppm) by following the procedure
described in ref 18. Details about its characterization are reported
in refs 18-20. The high purity silica sample (Aerosil) had a
volume) was followed by a molecular dynamics step (10 000
fs, 300 K, constant volume, constant energy). Finally, a last
minimization step was achieved (at constant cell volume). This
procedure has been applied to the pure silicalite structure and
to silicalites containing 11 HCl molecules/cell (simulating low
acid dosage) and 25 HCl molecules/cell (simulating high
dosage).
2
specific surface area of about 150 m /g. Before adsorption, the
silicalite and silica samples were activated at 973 K for 2 h.
The thermal treatments (under high vacuum) and the spectro-
scopic measurements have been performed on samples in the
form of self-supporting wafers.
High-purity HCl was produced by reaction of H2 with optical
AgCl at 973 K and used after elimination of residual H2 through
repeated freezing cycles, by using a cold trap kept at 77 K.
Ab initio calculations were performed at the ab initio level
2
3
by means of the GAUSSIAN 94 package. The gradient
corrected exchange and correlation hybrid functional B3-LYP24
with the 6-31+G(d,p) basis set was used for all calculations.
The geometry optimization was carried out by using analytical
gradient techniques. Harmonic normal-mode frequencies were
computed by adopting analytical second-energy derivatives and
solving the equations of nuclear motion by standard methods.25
(CH3)2O, (CH3CH2)2O, THF, and benzene were high-purity
Aldrich products; all the reagents were dosed from the gas phase.
The cell, equipped with a thermocouple in contact with the
sample holder and permanently attached to a gas manifold,

HCl and HCl-Base Adducts in Silicalite Channels
J. Phys. Chem. B, Vol. 102, No. 52, 1998 10755
Figure 1. IR spectra of silica (outgassed at 973 K) and of silica interacting with 4133 Pa of HCl at 300 K.
Figure 2. IR spectra of increasing doses of HCl adsorbed at 214 ( 5 K on silica outgassed at 973 K (0 e equilibrium pressure e 2933 Pa).
3
. Results and Discussion
silica sample (Aerosil) outgassed at T ) 973 K (Figure 1) is
-
1
characterized by a single narrow peak at 3750 cm due to the
stretching mode ν(OH) of the surface silanols; these SiOH
groups are isolated and not interacting each other, as indicated
by the very small fwhm (full width at half-maximum) of the
band.
The whole subject is divided in three parts (3.1-3.3). The
first part concerns the IR and theoretical study of the interaction
between HCl and silica and silicalite; this introductory study is
necessary, because data about the interaction of HCl with these
systems are not available in the literature (to our knowledge
the reader will find only ref 26 about the interaction of HCl
with NaY zeolite) and because it are of great help in under-
standing the spectroscopy of HCl-B adducts. The second part
is devoted to the study of the synthesis of HCl-B complexes
The IR spectra concerning the interaction of HCl (p ) 4133
Pa) with silica at 298 K are reported in Figure 1. It can be
immediately observed that the rotovibrational peaks of gaseous
-
1
HCl centered at 2886 cm appear above the unmodified
spectrum of silica; consequently, it can be deduced that at T )
(B ) base molecule) on silica and in silicalite channels and to
2
98 K neither chemical nor physical HCl adsorption on the
their spectroscopic properties. Finally, in the third part, the
results obtained for HCl-B adducts in silicalite are compared
to those obtained in cryogenic matrixes and in solution.
surface is taking place.
On the contrary, the IR spectra of increasing doses of HCl
(pmax ) 2933 Pa) at T ) 214 ( 5 K (Figure 2) show that the
adsorption is taking place for the following reasons.
3
.1. Adsorption of HCl on Silica and Silicalite. 3.1.1. IR
-
1
Results. The IR spectrum in the 3800-2200 cm range of the

1
0756 J. Phys. Chem. B, Vol. 102, No. 52, 1998
Paz e´ et al.
°C) (ν(HCl) ) 2780 cm ) and to those of (HCl)n oligomers
-
1 27
CHART 2
2
8-33
isolated in a cryogenic matrix (Ar or N2).
In the second
case, the interaction with the siloxane bridge is dominating;
unfortunately, the involved stretching frequencies cannot be
compared with those of suitable model compounds. For the time
being, it is not possible to make a reasonable choice between
the two possibilities. This problem will be discussed in the
following sections on the basis of molecular mechanics and ab
initio calculation.
1
(
i) The peak at 3750 cm- (νOH) of silanols is gradually
eroded with simultaneous formation, at lowest dosages, of three
bands at 3723, 3630, and 3580 cm . As the equilibrium
pressure of HCl increases, the intensity of the components at
723 and 3630 cm decreases in succession, while the intensity
of the band centered at 3580 cm grows in a parallel way.
Finally, at higher pressures, only a single band at 3595 cm is
observable. These results are a clear indication of the gradual
formation of hydrogen-bonded adducts SiOH(HCl)n (n g 1)
between the surface silanols and the HCl molecules (Chart 2).
About the detailed structure of the SiOH(HCl)n (n g 1)
clusters, it is not possible from the bare IR data to infer the
exact geometry of the individual complexes. However, helpful
information can be achieved by means of theoretical calculation,
and the results are shown later.
-
1
-1
3
iii) A band shows up at 923 cm-1 (inset Figure 2), whose
(
-
1
integrated intensity varies with HCl dosages in a way propor-
-
1
-1
tional to the intensity of the bands at 3594 and 2640 cm (due
to the complexes between HCl and silanols). This band is
tentatively assigned to the (in-plane) bending mode δ of the
silanol groups involved in the hydrogen-bonding interaction.
The spectrum of a silicalite sample outgassed at 973 K is
reported in Figure 3 (curve 1). It shows a single peak at 3746
-
1
cm ; with respect to silica, the peak is less intense and tailed.
This peak is due to silanols present on the external and internal
-
1
(
lower frequency tail, 3740-3680 cm ) surfaces of the
(
ii) Beside the narrow rotovibrational components of HCl gas
centered at 2886 cm , an absorption centered at 2788 cm
and a band (large and probably composite) centered at about
1
8,19
microcrystal.
-
1
-1
Like on silica, HCl is not adsorbed at room temperature (and
p = 2266 Pa); accordingly, the rotovibrational peaks of the gas
appear above the unmodified spectrum of the zeolite. Like on
silica, the absorption process takes place at T ) 214 K.
-
1
2
640 cm appear. The formation of the last two bands
demonstrates that, besides a dominant fraction of noninteracting
HCl molecules in the gas phase, at least two types of HCl
molecules perturbed in different way exists on the surface. Since
by increasing the HCl pressure a clear correlation is observed
between the bands at 2640 and 3594 cm , the band at 2640
cm is assigned to the ν(HCl) stretching mode in SiOH(HCl)n
n g 1) complexes (Chart 2).
The band at 2788 cm is continuously increasing with the
HCl pressure, even when the band at 2640 cm has reached a
constant maximum (that corresponds to the total consumption
of the band at 3750 cm due to free silanols). The band at
788 cm can be associated either with the formation of (HCl)n
oligomeric chains or with HCl species interacting with the
oxygen atoms of siloxane bridges. In the first case, each HCl
molecule is more stabilized by lateral interactions with contigu-
ous molecules than by interaction with the surface, and the IR
properties should be very similar to those of liquid HCl (at -100
In Figure 3, the interaction of increasing doses of HCl at T
=
214 K is shown. The following features can be noticed.
(1) The progressive disappearance of the peak at 3746 cm-1
-
1
-
1
(the tail at lower frequencies being preferentially affected) and
the simultaneous appearance of a broader band (without distinct
components, as on silica) varying with dosage between 3630
(
-
1
-
1
-1
(low pressures) and 3600 (high pressures) cm are evident.
These effects are clearly due to the formation of SiOH-(HCl)n
adducts between silanols (prevalently in internal positions) and
HCl molecules (Chart 2). HCl-OH adducts are more clearly
defined on amorphous silica since the silanol species on silica
are much more homogeneous (as shown by the very sharp peak
-
1
-1
2
-
1
at 3745 cm ), while a wider distribution of SiOH exists on
silicalite (the band at 3746 is asymmetric and less narrow) due
1
8,19
to the presence of hydroxyl nests.
Figure 3. IR spectra of increasing doses of HCl adsorbed at 214 ( 5 K on silicalite (outgassed at 973 K) (0 e equilibrium pressure e 2266 Pa).

HCl and HCl-Base Adducts in Silicalite Channels
J. Phys. Chem. B, Vol. 102, No. 52, 1998 10757
Figure 4. Molecular mechanics simulation of the adsorption of HCl in silicalite (11 molecules per cell), details in Experimental Section: (a) view
along [010] direction; (b) view along [100] direction; (c) view along [001] direction.
The shift of the band due to perturbed hydroxyls toward lower
frequencies as the dosage increases is presumably due to the
progressive formation of SiOH-(HCl)n complexes characterized
by increasing n. Considerations about the structure of the
SiOH(HCl)n (n g 1) clusters will be made in the next paragraph
on the basis of the ab initio results.
inside the channels of the zeolite; this assignment is based on
the data available for liquid HCl and for HCl entrapped in
2
7
2
8-33
cryogenic matrixes.
At low pressures, the main peak is
-1
-1
centered at 2830 cm and then it moves gradually to 2820 cm
by increasing the pressure. This shift is due to the formation of
oligomeric chains of HCl whose length is progressively longer;
as a matter of fact, IR data concerning (HCl)n complexes isolated
-
1
(2) A complex absorption between 3050 and 2350 cm
36,37
appears. This absorption shows a main asymmetric maximum
at 2820-2830 cm-1 (Q branch) flanked at higher and lower
frequencies by two weak and broad absorptions (at 2850-2800
in cryogenic matrixes of Ar and N2
show that the frequency
of the ν(HCl) mode decreases as the size (n) of the cluster
increases. The frequency of the Q branch maximum in silicalite
is higher than that of the same component on silica; this can be
explained by the presence, in the zeolite channels, of oligomeric
chains shorter than those on silica, the length of the chains being
limited by the channel structure environment (vide infra). The
frequency of the Q branch maximum is also lower than that of
-
1
and 2750-2650 cm ). By analogy with the spectra of CO
adsorbed on silicalite19 and of HCl at high pressures, these
components can be assigned to the P and R components of a
partially hindered rotator.34 The formation of a dense state of
HCl in the narrow channels is the characteristic result of the
van der Waals forces associated with the internal surface of the
pores. For the same temperature and pressure, this effect is not
present on external surfaces, where only the hydroxyl groups
are acting as adsorbing centers. An alternative explanation
27
-
1
the free gas (2886 cm ); this is due to the combined effect of
the high “pressure” of HCl in the channels and the interaction
with the zeolitic walls.
The IR manifestation of the complexes between HCl and
silanols, similar to those observed on silica (Figure 1), are
superimposed to the P envelope; this explains why the P branch
appears distinctly stronger and broader with respect to the R
branch.
35
reported in the literature for HCl in CCl4 is that they are
combination modes between the ν HCl mode (Q branch) and
other (not well defined) low-frequency intermolecular solute-
solvent modes. In the following description, the PQR nomen-
clature is used. Notice how a broader adsorption is superimposed
to the P branch.
As already mentioned, the HCl molecules physisorbed inside
the channels show very peculiar rotovibrational features. Both
the Q branch (corresponding to a forbidden transition in the
On the top of these absorptions, the rotovibrational peaks of
-1
27
gaseous HCl (centered at 2886 cm ) are clearly distinguishable.
The presence of PQR envelopes (without rotovibrational
components) is clearly distinguishing the spectrum of HCl
adsorbed into the channels from that adsorbed on external
surfaces (Aerosil).
gas but active in the liquid ) and the Q and R branches (well
defined in the gas but not resolved into the individual compo-
nents when the rotation of the molecules is partially hindered,
as it happens for example at very high pressures of gas or in
3
4,35,38
solution
) are present. These peculiarities show that a
The absorption at 2820-2830 cm- is due to HCl physisorbed
1
fraction of the physisorbed molecules is in a state similar to

10758 J. Phys. Chem. B, Vol. 102, No. 52, 1998
Paz e´ et al.
Figure 5. Molecular mechanics simulation of the adsorption of HCl in silicalite (25 molecules per cell), details in Experimental Section: (a) view
along [010] direction; (b) view along [100] direction; (c) view along [001] direction.
that of a dense gas (in which the intermolecular forces are still
strong); in this state, the molecules are preserving a certain
rotational degree of freedom and behave therefore as hindered
rotators. In addition, the presence of a strong Q central
component shows that they are in equilibrium with a more dense
phase similar to the liquid.
The presence of the P, Q, and R branches only for HCl
adsorbed on silicalite (and not on silica) demonstrates that in
the zeolitic cavities the adsorbed HCl molecules maintain some
rotational freedom and that the same effect cannot be obtained
on external surfaces.
To understand how HCl molecules are physisorbed inside
the zeolite channels, a qualitative procedure based on molecular
mechanics simulation of a low (11 molecules per cell) and a
high (25 molecules per cell) filling of the channels of a perfect
silicalite, i.e., a silicalite without internal defects (hydroxyl
nests), has been carried out. The energy-minimized structures
show that, at low dosage (Figure 4), the tendency to form
oligomeric species is very modest and that only at the highest
filling (Figure 5) a low density, liquid-like state containing a
large amount of HCl‚‚‚HCl hydrogen bonds is present. A
qualitative picture is then that at low coverage the interaction
energy between HCl molecules is smaller or very close to the
interaction energy between HCl and the zeolite walls. Obviously,
when highest filling conditions are reached, the HCl molecules
fill up the zeolite channels making long H-bonded strings. The
weakness of the HCl-HCl and HCl-walls interactions is also
in agreement with the presence of rotovibrational envelopes in
the IR spectra.
volume, it is not adequate for describing the interaction with
silanols, responsible for the erosion of the free silanol peak. To
shed some light on this topic, ab initio methods are a more
convenient choice.
As a model of the adsorption site, the largely used minimal
cluster model H3SiOH (structure S, Figure 6a) has been
3
9
adopted. Six hypothetic structures of SiOH(HCl)n (n g 1)
clusters (Figure 6a) have been chosen to simulate different
dosages corresponding to different experimental equilibrium
pressures. The results can be summarized as follows. (i) For
the 1:1 HCl-SiOH adduct, the only stable adduct is shown in
Figure 6a as SA1, in which the silanol behaves as a hydrogen
bond acceptor with respect to HCl. Other possible 1:1 structures,
like the cyclic one or that in which S acts as a hydrogen bond
donor, resulted unstable on the potential energy surface. (ii) As
far as the 2:1 HCl-SiOH stoichiometry is considered, two
configurations have been studied among all the possible ones:
a cyclic structure (SR2, Figure 6a) and a structure in which S
acts as a double hydrogen-bond acceptor (SA2, Figure 6a). As
far as the 3:1 stoichiometry, a cyclic complex (SR3, Figure 6a)
as well as a structure in which S acts as both a hydrogen-bond
acceptor and donor with respect to HCl (SA3, Figure 6a) have
been considered. Finally, the siloxane-HCl interaction has been
taken into account (SXA, Figure 6a) by mimicking the siloxane
bridge with the (SiH3)2O molecule SX.
The ab initio study has also been extended to aggregates made
by the self-interaction of HCl molecules by considering clusters
spanning from (HCl)2 to (HCl)4, respectively. Five HCl self-
aggregate structures have been computed, as shown in Figure
6
b.
For all structures shown in parts a and b of Figure 6, the
optimized geometries and the full set of harmonic frequencies,
3
.1.2. Ab Initio Calculation. While molecular mechanics
simulations can give an insight into the HCl-HCl and HCl-
walls interaction in hydroxyl-free portions of the internal

HCl and HCl-Base Adducts in Silicalite Channels
J. Phys. Chem. B, Vol. 102, No. 52, 1998 10759
Figure 6. (a) Molecular drawings for silanol and siloxane adducts with HCl. For labeling the selected complexes, the following acronyms are
used: S ) silanol, SX ) siloxane, A ) acceptor, R ) ring. Each HCl molecule has been labeled to help the discussion (see Table 1). Intermolecular
bond distances are in angstroms. (b) Molecular drawings for HCl self-aggregates. Adducts labels assigned as L ) linear adducts and R ) ring
adducts. Each HCl molecule has been labeled to help the discussion (see Table 1). Intermolecular bond distances in angstroms.
as well as the interaction energies have been computed. A
selection of the results obtained is reported in Table 1.
The H3SiOH molecule does not exist as such because it is
unstable with respect to the condensation to disiloxane; however,
the calculated geometry (see Figure 6a) is in agreement with
previously published data computed at higher levels of calcula-
tion.40
The geometries of the HCl self-aggregates (see Figure 6b)
show intermolecular H‚‚‚Cl distances well within a 2.5 Å limit,
with some modulation in the ring structures (R3 and R4) due
to cooperative effects.
For each adduct, the interaction energy ∆En has been
computed from the total energies of each molecular constituents
of the adduct, by the formula
The analysis of the geometries of silanol-HCl and siloxane-
HCl adducts shows a number of interesting features. (i) The
HCl molecule acts preferably as a proton donor, being, at the
same time, a poor proton acceptor. Indeed, the intermolecular
distances of the SR2 and SR3 adducts reveal a noticeably shorter
H-O bond with respect to the SA1 adduct between the terminal
HCl molecule and the basic oxygen atom belonging to S (1.749
and 1.665 Å for SR2 and SR3, respectively). Also worth noting
is the shrinking of all intermolecular distances when passing
from the SR2 to the SR3 adduct, a clear indication that
cooperative effects are acting along the ring of H-bonded
molecules. (ii) The Cl‚‚‚Cl repulsion is responsible for the
lengthening of the H‚‚‚O bond in the SA2 adduct when
compared with SA1. (iii) The hydrogen bond between HCl and
the siloxane bridge (structure SXA in Figure 6a) is noticeably
short, a fact almost unexpected because of the very weak basicity
of the oxygen atom belonging to the siloxane bridge.
∆
E ) E(A(HCl) ) - E(A) - nE(HCl)
n n
in which A is the molecular partner (either S, SX, or HCl) of
the HCl molecule in the corresponding aggregate A(HCl)n and
n is the total number of HCl molecules involved. The interaction
energy per HCl molecule is defined as ∆E1 ) ∆En/n, from
which the corresponding heat of formation at 0 K is computed
by adding to both ∆En and ∆E1 the proper zero-point energy
correction to get ∆H n(0) and ∆H 1(0), respectively.
The analysis of data in Table 1 shows the following trend in
the magnitude of ∆H n(0) for all adducts:
0
0
0
L2 < L3 < R3 = SXA = L4 < SA1 = R4 < SA2 <
SR2 < SA3 < SR3
To compare the heats of formation on a common gauge, we
have reported in Figure 7 the heats of formation per HCl

10760 J. Phys. Chem. B, Vol. 102, No. 52, 1998
Paz e´ et al.
TABLE 1: Interaction Energy, ∆E
n
and Heat of Formation
0
at 0 K, ∆H
n
(0) (in kJ/mol) and Selected Unscaled Harmonic
Frequencies (in cm ) for All Adductsa
-
1
systema -∆Enb -∆H n(0)
0
b
ω(OH)c
3903
ω(HCl)c,d
HCl
S
2953
L2
6.7
2.6
2881 (-72)
946 (-7)
1
2
2
R3
21.5
11.1
2816 (-137) 1, 2, 3 sym
2
2
842 (-111) 2, 3 asym
845 (-108) 1-2, 3 asym
L3
R4
15.2
38.5
6.8
2845 (-108) 2
2
2
869 (-84)
941 (-12)
1
3
21.5
2737 (-216) 1, 2, 3, 4 sym
2
2
2
779 (-174) 1, 3 asym
779 (-174) 2, 4 asym
796 (-157) 1, 3-2, 4 asym
L4
24.4
11.9
2814 (-139) 2, 3 sym
2
2
2
835 (-118) 2, 3 asym
861 (-92)
943 (-10)
1
4
SA1
SR2
26.4
45.2
19.7
32.5
3879 (-24) 2669 (-284)
3797 (-106) 2504 (-449) 1
782 (-171) 2
3839 (-64) 2741 (-212) sym
774 (-179) asym
2
SA2
SR3
38.6
66.0
26.0
47.7
2
Figure 8. Experimental IR spectra of HCl adsorbed on silica at 214
( 5 K (p ) 100 Pa) (curve a). Fitting obtained using, as base
components, bands with Lorentzian shape centered at the calculated
frequency (corrected by a scaling factor of 0.9608) (curve b); the fwhm
of the components are obtained by the Pimentel equation that correlates
3719 (-184) 2336 (-617) 3
2
2
672 (-281) 2
738 (-215) 1
SA3
52.1
18.0
35.0
11.6
3727 (-176) 2709 (-244) 2, 3 asym
2
2
749 (-204) 2, 3 sym
933 (-20)
-1 7
FWHM and ∆ νj shifts of frequency (FWHM ) 0.72∆ νj + 2.5 cm ).
1
SXA
2720 (-233)
aggregation bewteen HCl and silanol sites, a fact resulting from
the poor proton acceptor capability of HCl itself.
a
Adopted labels accordingly to Figure 6. b Uncorrected for BSSE.
c
Corresponding frequency shifts with respect to the unperturbed
d
Hence, the possible steps of the HCl adsorption are the
formation of the complexes with hydroxyl groups of the surface
and, when all the silanols are consumed, the interaction of HCl
with the siloxane bridges. Only after this stage does self-
aggregation of HCl start to take place.
frequencies are enclosed in parentheses. Attribution of each ω(HCl)
frequency to the corresponding HCl molecule following the labelling
scheme of Figure 6.
Table 1 also includes a selection of the computed harmonic
vibrational frequencies as well as the corresponding shifts caused
by the H-bonding interaction. The reported frequencies are for
the OH and HCl stretching modes. The analysis of the data
reported in Table 1 reveals that the largest ω(HCl) shifts are
for the cyclic structures, namely R3, R4, SR2, and SR3, in which
shifts almost twice as large as those computed for the corre-
sponding noncyclic structures are computed. This fact, correlates
with the higher interaction energies for such structures and is
indeed a measure of the H-bonding cooperativity taking place
in the cyclic adducts.
In Figure 8, the experimental spectrum of HCl adsorbed on
silica, as far as concerns the OH stretching region, is interpreted
on the basis of the ab initio calculations. A fit of the
experimental spectrum has been made by using, as base
components, bands with Lorentzian shape centered at the
frequencies calculated theoretically (corrected by a scaling factor
of 0.9608). The adopted FWHM of the components were
obtained from the Pimentel equation that correlates FWHM and
Figure 7. Bar chart of the heat of formation per HCl molecule,
0
-
∆H
1
(0), for all adducts. See Figure 6 for the labeling scheme.
Energetic data in kilojoules per mole.
-1 7
the frequency shifts ∆ νj (FWHM ) 0.72∆ νj + 2.5 cm ). The
intensities of the components were obtained through a fitting
0
-1
molecule at 0 K, -∆H 1(0). A monotonic increase of its value
procedure. The experimental band at 3723 cm , present only
is seen for the HCl self-aggregates. Values much higher than
the maximum value of 6.5 kJ/mol computed for R4 are seen
for the aggregates involving HCl with either silanol or siloxane
molecules. Among these latter, the SA1 shows the largest value
of -∆H 1(0) (19 kJ/mol), followed by the two ring structures
SR2 and SR3 (around 16 kJ/mol). These data show that the
self-aggregation of HCl is less favored than processes involving
at the lowest pressures, is therefore assigned to the formation
-
1
of a HCl-SiOH 1:1 adduct; the component at 3630 cm is
assigned to 2:1 complexes, while the absorption at about 3600
-
1
cm is due to complexes with 3:1 (or more) stoichiometry.
When a simulation of the spectra of silica at high HCl
loadings was performed, not so good results were obtained. In
0
-
1
particular, the components in the 3620-3570 cm range are

HCl and HCl-Base Adducts in Silicalite Channels
J. Phys. Chem. B, Vol. 102, No. 52, 1998 10761
missing. This could be explained by two facts: (1) siloxane-
HCl complexes are stable, and their population increases
preferentially as the pressure increases; moreover the fit does
not consider that the presence of HCl molecules adsorbed on
the siloxanes, surrounding the SiOH-HCl adducts, is certainly
influencing the stretching frequency of the silanol and (2) other
adduct structures are possible at high pressure, due to the
formation of multilayer HCl adsorption on silica-HCl and
siloxane-HCl adducts.
TABLE 2: Experimental Frequencies of ν(HCl) Mode in
HCl-B Complexes in Silicalite at 214 ( 5 Ka
benzene CH3CN (CH3)2O (CH3CH2)2O
THF
νj HCl (cm-1
304a
∆
)
111
526 ( 20
556 ( 20
606 ( 20
CHART 3
the other complexes (with benzene, diethyl ether, and THF) are
summarized in Table 2.
In principle, a simulation of the spectra in the ν(HCl)
stretching region of frequencies could also be made. To this
aim, all the frequencies of the collective modes of HCl-silanol
complexes, all the collective modes of HCl clusters, and even
the mode of the siloxane-HCl adduct should be considered.
However, the number of base components needed to simulate
the experimental spectrum would be exceedingly large (more
than 20 bands), and this would lead to an arbitrary fit. Without
describing the matter in detail, it must be remarked that, despite
all the difficulties mentioned before, all the computed frequen-
cies shifts lie in the region where the absorption is experimen-
tally observed. In particular, the HCl adducts’ calculated
3. 2. 1. Silicalite-(CH3)2O and Silica-(CH3)2O. To facilitate
the discussion of the spectra of the HCl-(CH3)2O complexes
formed in the silicalite channels (vide infra), a preliminary
analysis of the spectra of the silicalite-ether system (i.e., without
HCl) is useful (Figure 9). The spectra obtained on Aerosil are
very similar (only the intensity is lower), hence they are not
shown for the sake of brevity.
As the pressure of (CH3)2O increases, the band at 3746 cm-
1
is completely eroded; simultaneously, a broader band shows
-
1
up at 3246 cm . A relation of strict proportionality exists
between the decrease of the integrated intensity of the peak of
-
1
frequency shifts lie in the range 50-180 cm , while the
-
1
free silanols (3746 cm ) and the increase of the integrated
frequency shifts of the silanol and siloxane adducts are over
-
1
-
1
intensity of the band centered at 3294 cm ; moreover, a clear
1
80 cm , again confirming the previous interpretation.
-
1
isosbestic point appears at 3580 cm . These two facts
demonstrate that a 1:1 complex is gradually formed (Chart 3).
The band of the ν(OH) mode of the silanol groups is shifted
All the considerations made about silica can be extended to
silicalite. In the latter case, however, the surface is prevalently
internal so that complexes with size greater than three, are less
probable. This consideration explains why the absorption in the
-
1
-1
downward by about 500 cm (FWHM = 300 cm ), a figure
typical of a hydrogen bond of medium strength.
-
1
3
700-3600 cm range has its maximum at frequency values
The bands of the methyl stretching modes (unaffected by the
hydrogen-bonding interaction) are observed in the 3000-2700
higher than those of silica (because the steric constrains favor
the prevalent formation of complexes with n ) 2). This
conclusion is confirmed by a fit (not reported for the sake of
-
1
-1
cm interval (νsym ) 2828, 2811 cm , νasym ) 2989, 2945-
-
1
2
866 cm ); these bands are made complicated by the Fermi
-
1
brevity) in the 3800-3500 cm range; in fact, while the
components are the same as those observed on silica, the
intensity distribution is different because the 3:1 complexes are
present in lower proportion. Furthermore, in the case of HCl
adducts (ν(HCl) region of frequency), the IR band is observed
at higher frequency than in silica. That trend could be explained
on grounds of the linear HCl adducts computed frequency shifts.
Indeed, such complexes, which have lower steric constraints
and lower frequency shifts, are stabilized in the channels of
zeolite, with respect to the other bulky configurations.
interaction with the two δ(CH3) modes. The corresponding
-
1
bending modes are at 1473 and 1455 cm . Also, these bands
are not influenced by the adsorption state. Finally, the bands
due to the CH3 rock mode are observed at 909 (hydrogen
-1
bonded) and 922 (physisorbed) cm . Notice how, in the spectra
corresponding to highest equilibrium pressures (dashed lines
in Figure 9), many very weak bands are observable (at 2694,
2
1
634, 2627, 2593, 2562, 2536, 2406, 2250, 2162, 2080, 2006,
839, and 1586 cm ) due to combination and overtones of the
-
1
ether molecule.
. 2. 2. Silicalite-HCl-(CH3)2O. The spectra of increasing
Even if a surprisingly good correlation between experimental
and theoretical results is verified, a few considerations about
the calculation method must still be made. (1) The calculated
frequencies are harmonic; anharmonic contributions are therefore
missing. (2) DFT methods tend to overestimate the frequency
shift in hydrogen-bonded systems; moreover, the shift is further
3
doses of (CH3)2O on the silicalite-HCl system (pHCl ) 1467
Pa) at T ) 214 ( 5 K are shown in Figure 10.
The following observations were made.
-1
(
i) The peak of free silanols at 3750 cm and the band
-1
centered at about 3600 cm (due to SiOH(HCl)n) are gradually
41
overestimated in the case of multiple interactions. This fact
may be due to the presence of the self-interaction error, which
is more significant in systems (like those here considered) in
which protons are involved.42 (3) The overestimation of the
frequency shifts is also due to the effect of temperature, because
the calculated shifts were computed at 0 K while the experiment
was made at 214 ( 5 K. (4) Finally, it is known that the minimal
cluster model chosen to mimic the isolated hydroxyls of the
surface (structure S, Figure 6) is less acidic than the real surface
eroded upon (CH3)2O dosage. Simultaneously, the broad band
-1
(
centered at 3242 cm ) due to the (CH3)2O-silanols adducts
appears. These modifications are due to the reactions shown in
Chart 4. From this it is inferred that the HCl-silanols complexes
are less stable than those formed between ether and silanols, in
agreement with the corresponding shifts ∆ νj of the ν(OH) modes
of silanols.
-1
(ii) The peak at 2830 cm , characteristic of HCl adsorbed
on silicalite (bold curve in Figure 10), gradually desappears with
simultaneous formation of a broad band centered at about 2360
4
0
SiOH group. As a result of this, the O atom is more basic and
this in turn favors the preferential formation of adducts where
the HCl is acting as a proton donor.
-1
-1
cm (FWHM ≈ 530 cm ). On the analogy with the known
43,44
data concerning HCl-(CH3)2O complexes in gas
solid phase, the absorption at 2360 cm is ascribed to the
and in the
4
5
-1
3
. 2. HCl-Base Complexes. For the sake of brevity only
the spectra of the adducts formed with dimethyl ether will be
illustrated in detail. As for the spectra obtained with CH3CN,
the reader is referred to ref 17. The relevant data concerning
ν(HCl) mode in the 1:1 HCl-(CH3)2O adducts (∆ νj (HCl) =
-1
540 cm ), formed according to the equilibrium reaction shown
in Chart 5.

10762 J. Phys. Chem. B, Vol. 102, No. 52, 1998
Paz e´ et al.
3 2
Figure 9. IR spectra of increasing doses of (CH ) O adsorbed on silicalite at 214 ( 5 K (0 e equilibrium pressure e 13300 Pa).
3 2
Figure 10. IR spectra showing the interaction of HCl adsorbed on silicalite (1467 Pa) with increasing doses of (CH ) O at 214 ( 5 K (see text
for details).
CHART 4
Like in the gas phase, the proton transfer does not occur44
and the adduct has neutral character and is characterized by a
medium hydrogen bond. The large band is made complicated
in the 2200-1500 cm-1 range by several Fermi resonance
effects due to coupling of the ν(HCl) mode with the [δ(O-
4
4
HCl) + ν(COC)] combination modes. The intensity of this
absorption is much higher than that of the same complex formed
in an analogous experiment on silica (data non shown for

HCl and HCl-Base Adducts in Silicalite Channels
J. Phys. Chem. B, Vol. 102, No. 52, 1998 10763
CHART 5
brevity); this confirms that the 1:1 complexes are prevalently
formed in the zeolite channels. It must be stressed that the
ν(HCl) stretching mode of (CH3)2O-HCl adducts is at 2570
-1
43,44
-1
45
cm in the gas phase
and at 2530 cm in the solid phase.
(
iii) The stretching modes of methyl groups are observed in
-
1
the 3100-2800 cm range (νasym ) 2992, 2936, 2892, 2875
-
1
-1
cm ; νsym ) 2826, 2820 cm ), while the bending modes are
-1
-1
in the 1500-1400 cm range (1473, 1457, 1426 cm ). Finally,
the rocking modes are at 905 (hydrogen bonded) and at 921
-1
(
physisorbed) cm . Most of these frequencies are slightly higher
Figure 11. BHW plot of the shifts (∆ νj ) of the ν(HCl) frequencies in
1:1 ClH-B hydrogen-bonded complexes with different B in different
environments versus the shifts (∆ νj ) of SiOH groups in 1:1 hydrogen-
bonded complexes with the same bases. 0 represents the experimental
than those obtained in the absence of HCl; this also indicates
that the modes of the methyl groups are slightly perturbed by
the interaction with the HCl.46
28,39,52
28
28
data of HCl in an Ar criogenic matrix (B ) N
2
,
CO
2
,
CO,
3. 2. 3. HCl-B (B ) (CH3CH2)2O, THF, Benzene) Complexes
28
28
66
C
2
H
4
,
H O, CH CN, ); O represents the experimental data HCl
2 3
in Silicalite. The study of the interaction of HCl with basic
molecules has been extended also to other bases like benzene,
diethyl ether, and THF. The main features of the spectra are
similar to those obtained for the interaction with (CH3)2O and
are summarized in Table 2, where the shifts ∆ νj (HCl) induced
by the interaction with the base B are reported. The FWHM of
the bands are about (3/4)∆ νj as usually observed in hydrogen-
bonded systems.
inserted in silicalite; 4 represents the experimental data of HCl in a
solution of apolar solvents (B ) CCl , chloroform, C Cl H , benzene,
4
2
2
2
nitrobenzene, toluene, anisole, p-xylene, mesitylene, ethyl acetate,
61
acetone, and dioxan ).
of dipoles on the atoms of the “solvent” (cryogenic matrix,
zeolite, and solution); in fact, the electric field induces an
increase of the ionic character of the hydrogen bond (by
changing the effectiveness of the screening effects), with a
subsequent increase of ∆ νj . This effect is well-known in solution
as “solvent effect” and in cryogenic matrixes as “matrix effect”.
In the macroscopic scale, it is connected to the particular
dielectric constants of the media where the complexes are
3
. 3. BHW Plot. A well accepted method to study by means
of IR spectroscopy the acid strength of a Brønsted acid HX is
based on the evaluation of the frequency shifts that the ν(X-
H) mode undergoes by interaction with molecules of different
3
-6
basicity. When these shifts are plotted against the shifts that
a second acid molecule ZH (considered as reference) is
undergoing by interaction with the same bases (∆ νj (HX) vs
formed as discussed in a pioneer paper by Kirkwood.65
A
quantitative analysis in terms of the Kirkwood equation would
require the definition of the dielectric constant of a nonhomo-
geneous medium; for this reason, it is not considered.
It is so concluded that the effect of the solid environment
represented by the neutral silicalite framework on the spectro-
scopic properties of the HCl-B adducts entrapped in the
channels is relevant (it is intermediate between that observed
in apolar solvents and in an argon matrix), and consequently
plays a role in determining the strength of the Brønsted groups
located in internal positions. This conclusion, verified for the
neutral framework of silicalite (where the interactions between
the acid guest molecule and the zeolite host walls are prevalently
of the van der Waals type), is a fortiori valid for the much
stronger structural Al(OH)Si Brønsted groups of the zeolites.
However, to quantitatively extend these considerations to real
zeolites, the spectroscopy of HCl-B model complexes formed
in situ into charged frameworks such as those of H-ZSM-5,
H-Y, etc. should be investigated by performing a new series
of experiments.
∆
νj (HZ)), the Bellamy, Hallam, Williams (BHW) plot is
11-13
obtained.
It has been found that for each studied acid
molecule a single straight line is observed whose gradient is a
measure of the acidity of the molecule. It has been shown that
the BHW plot can be usefully employed also to estimate the
acid strength of Brønsted groups located on surfaces (provided
1
0,16
that the proton affinity of the base B is not too high).
It has
been experimentally verified that a good reference acid group
for heterogeneous systems is the SiOH group of silica and of
1
0,16
silicalite.
It is useful to recall that the HCl molecule has been
4
7-51
extensively studied in the past, in the gaseous phase,
in
52-60
61-64
cryogenic matrixes,
and in solution,
and that numerous
data sets about the ν(HCl) frequencies in various HCl-B
complexes can be found in the literature. If all the experimental
data are reported in a BHW plot, Figure 11 is obtained. In this
figure, the ∆ νj (HCl) values of the intrazeolitic HCl-B com-
plexes (Table 2) are also reported for comparison.
From Figure 11, it is clearly emerging that the data of the
HCl complexes in the three different environments (in silicalite,
matrix, and apolar solvents) originate three different straight
lines. Unfortunately, to our knowledge, not so many data about
HCl-B complexes in an Ar matrix are available, and line fitting
the cryogenic data is certainly not accurate; despite this, there
is no doubt that the slope of the HCl-B complexes encapsulated
in silicalite is greater than that of HCl-B complexes isolated
in a cryogenic matrix and in vacuo. The general trend of the
slopes of the three straight lines (that is, m(HCl in solvent) >
m(HCl in silicalite) > m(HCl in Ar matrix)) can be explained
in terms of the electric field effects associated with the presence
4
. Conclusions
The IR study of the HCl-silica and HCl-silicalite interaction
at ∼210 K indicates that a weak H-bond interaction occurs
between silanols and HCl molecules and that adducts of
increasing nuclearity (containing two or more HCl molecules)
are formed upon increasing the HCl dosage. In silicalite, the
HCl molecules interact with the siloxane bridges and form also
small (HCl)n clusters. The molecules behave as partially
hindered. A complete assignment is given through the help of
molecular mechanics and ab initio calculations. The interaction
of adsorbed HCl with basic molecules B [(CH3)2O, (CH3CH2)2O,

10764 J. Phys. Chem. B, Vol. 102, No. 52, 1998
Paz e´ et al.
THF, and benzene)] leads to the formation of intrazeolitic 1:1
adducts HCl-B. The comparison of the ∆ νj (HCl) values of
various HCl-B complexes in different environments (in apolar
solvents and in an Ar cryogenic matrix) with the corresponding
values of HCl-B intrazeolitic complexes demonstrates that the
acidity of HCl grows substantially on passing from the cryogenic
matrix, to silicalite, and to solution. It is concluded that the
zeolitic framework of the zeolite acts as a solvent that influences
the acidity of structural Si(OH)Al Brønsted sites located in the
channels and cavities. The contribution of the tridimensional
structure of the zeolitic framework in determining the acid
strength of the Si(OH)Al groups cannot consequently be ignored.
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 94, revision C.3; Gaussian,
Inc.: Pittsburgh, PA, 1995.
(
24) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. Lee, C.; Yang, W.;
Parr, R. C. Phys. ReV. B 1988, 37, 75.
25) Wilson, E. B.; Decius, J. B.; Cross, P. C. Molecular Vibration;
(
McGraw-Hill: New York, 1955. Hehre, W. J.; Radom, L.; Schleyer, P. V.
R.; Pople, J. A.; Ab initio Molecular Orbital Theory; Wiley: New York,
1
986.
(26) Ozin, G. A.; O¨ zkar, S.; Stucky, G. D. J. Phys. Chem. 1990, 94,
7562. Ozin, G. A.; O¨ zkar, S.; McMurray, L. J. Phys. Chem. 1990, 94, 8289;
1990, 94, 8297.
(
(
27) West, W. J. Chem. Phys. 1939, 7, 795.
28) Barnes, A. J.; Hallam, H. E.; Scrimshaw, G. F. Trans. Faraday
Soc. 1969, 65, 3172.
29) Pine, A. S.; Howard, B. J. J. Chem. Phys. 1986, 84, 591.
(30) Girardet, C.; Robert, D. J. Chem. Phys. 1973, 58, 4110.
Acknowledgment. The authors thank Prof. Piero Ugliengo
and Dott. Gabriele Ricchiardi for fruitful discussion and wise
advice and Dott. Giuseppina Meligrana for help in IR measure-
ments. This work has been supported by CNR (Progetto
Strategico Tecnologie Chimiche) and MURST.
(
(31) Andrews, L. Faraday Discuss. Chem. Soc. 1988, 86, 37.
(32) Girardet, C.; Robert, D. J. Chem. Phys. 1973, 59, 5020.
(33) Hallam, H. E. In The Hydrogen Bond; Schuster, P., Zundel G.,
Sandorfy C., Eds.; North-Holland, 1976; p 1075.
34) Jeremy Jones, W. In Infrared Spectroscopy and Molecular Struc-
ture; Davies, M., Ed.; Elsevier Publishing Company: Amsterdam, 1963; p
(
References and Notes
1
5, 159.
(
1) van Santen, R. A. Catal. Today 1997, 38, 377.
(35) West, W.; Edward, R. T. J. Chem. Phys. 1937, 5, 14.
(36) Maillard, D.; Schriver, A.; Perchard, J. P.; Girardet, C. J. Chem.
(2) Leopold, K. R.; Canagaratna, M.; Phillips, J. A. Acc. Chem. Res.
1
997, 2, 57. Tomasi, J.; Persico, M. Chem. ReV. 1994, 94, 2027.
Phys. 1979, 71, 505.
(37) Maillard, D.; Schriver, A.; Perchard, J. P.; Girardet, C.; Robert, D.
J. Chem. Phys. 1977, 67, 39.
(
3) Corma, A. Chem. ReV. 1995, 95, 559.
(4) Ugliengo, P.; Ferrari, A. M.; Zecchina, A.; Garrone, E. J. Phys.
Chem. 1996, 100, 3632.
(38) Herzberg, G. Molecular Spectra and Molecular Structure, I.
(
(
5) Zecchina, A.; Otero Are a´ n, C. Chem. Soc. ReV. 1996, 187.
Molecular Spectra of Diatomic Molecules; D. Van Nostrand Company,
Inc.: New York, 1945; p 464.
6) Sauer, J.; Ugliengo, P.; Garrone E.; Saunders, V. R. Chem. ReV.
1
994, 94, 2095.
(
39) Ugliengo, P.; Garrone, E. J. Mol. Catal. 1989, 54, 439
(7) Pimentel, G. C.; McLellan, A. L. The Hydrogen Bond; Pauling,
(40) Civalleri, B.; Garrone, E.; Ugliengo, P. J. Phys. Chem. B 1998,
L., Ed.; W. H. Freeman: San Francisco, 1960.
1
02, 2373.
41) Ugliengo, P.; Ferrari, A. M.; Civalleri, B.; Garrone, E. In prepara-
tion.
(
(
8) Kubelkov a` , L.; Beran, S.; Lercher, J. Zeolites 1989, 9, 539.
9) Zecchina, A.; Otero Are a´ n, C. Chem. Soc. ReV. 1996, 188. Bordiga,
(
S.; Lamberti, C.; Geobaldo, F.; Zecchina, A. Langmuir 1995, 11, 527.
Zecchina, A.; Bordiga, S.; Lamberti, C.; Spoto, G.; Carnelli, L.; Otero Arean,
C. J. Phys. Chem. 1994, 98, 9577.
(
42) Civalleri, B.; Garrone, E.; Ugliengo, P. J. Mol. Struct. (THEOCHEM)
1
997, 419, 227.
(
(
(
(
43) Bertie, J. E.; Falk, M. V. Can. J. Chem. 1973, 51, 1713.
44) Bertie, J. E.; Millen, D. J. J. Chem. Soc. 1965, 497.
45) Seel, R. M.; Sheppard, N. Spectrochim. Acta 1969, 25A, 1287.
46) Pelmenschikov, A. G.; Morosi, G.; Gamba, A. J. Phys. Chem. 1997,
(
10) Zecchina, A.; Bordiga, S.; Spoto, G.; Scarano, D.; Span o` , G.;
Geobaldo, F. J. Chem. Soc., Faraday Trans. 1996, 92, 4863.
11) Bellamy, L. J.; Hallam, H. E.; Williams, R. L. Trans. Faraday Soc.
958, 54, 1120.
(
1
1
01, 1178. Pelmenschikov, A. G.; Morosi, G.; Gamba, A. J. Phys. Chem.
(
(
(
12) Bellamy, L. J.; Pace, R. J. Spectrochim. Acta 1969, 25A, 319.
13) Hair, M. L.; Hertl, W. J. Phys. Chem. 1970, 74, 91.
14) Yin, F.; Blummenfeld, A. L.; Fripiat, J. J. J. Phys. Chem. B 1997,
1
992, 96, 2241. Pelmenschikov, A. G.; Morosi, G.; Gamba, A.; Zecchina,
A.; Bordiga, S.; Paukshtis, E. A. J. Phys. Chem. 1993, 97, 11979.
(
47) Lovejoy, C. M.; Nesbitt, D. J. Chem. Phys. Lett. 1988, 147, 490.
1
01, 1824. Zecchina, A.; Marchese, L.; Bordiga, S.; Paz e` , C.; Gianotti, E.
(48) Maillard, D.; Schriver, A.; Perchard, J. P.; Girardet, C.; Robert, D.
J. Phys. Chem. B 1997, 101, 10128.
15) Buzzoni, R.; Bordiga, S.; Ricchiardi, G.; Lamberti, C.; Zecchina,
J. Chem. Phys. 1977, 67, 3917.
(
(
49) Bertie, J. E.; Falk, M. V. Can. J. Chem. 1973, 51, 1713.
A.; Bellussi, G. Langmuir 1996, 12, 930. Datka, J. J. Chem. Soc., Faraday
Trans. 1 1981, 77, 2877. Datka, J.; Tuznik, E. J. Catal. 1986, 43, 102.
Datka, J.; Turek, A. M.; Jehng, J. M.; Wachs, I. E. J. Catal. 1992, 135,
(50) Sharpe, S. W.; Zeng, Y. P.; Wittig, C.; Beaudet, R. A. J. Chem.
Phys. 1990, 92, 943.
(51) Seel, R. M.; Sheppard, N. Spectrochim. Acta 1969, 25A, 1287.
(52) Andrews, L.; Davis, S. R. J. Chem. Phys. 1985, 83, 4983.
(53) Ault, B. S.; Pimentel, G. C. J. Phys. Chem. 1973, 77, 57.
(54) Ault, B. S.; Pimentel, G. C. J. Phys. Chem. 1973, 77, 1649.
(55) Johnson, G. L.; Andrews, L. J. Phys. Chem. 1983, 87, 1852.
(56) Sharpe, S. W.; Zeng, Y. P.; Wittig, C.; Beaudet, R. A. J. Chem.
2
0186. Emeis, C. A. J. Catal. 1993, 141, 347.
16) Paz e` , C.; Bordiga, S.; Lamberti, C.; Salvalaggio, M.; Zecchina,
A.: Bellussi, G. J. Phys. Chem. B 1997, 101, 4740.
17) Paz e´ , C.; Bordiga, S.; Spoto, G.; Lamberti, C.; Zecchina, A. J. Chem
Soc., Faraday Trans. 1998, 94, 309.
18) Zecchina, A.; Bordiga, S.; Spoto, G.; Marchese, L.; Petrini, G.;
Leofanti, G.; Padovan, M. J. Phys. Chem. 1992, 96, 4985.
19) Zecchina, A.; Bordiga, S.; Spoto, G.; Marchese, L.; Petrini, G.;
Leofanti, G.; Padovan, M. J. Phys. Chem. 1992, 96, 4991.
20) Marra, G. L.; Tozzola, G.; Leofanti, G.; Padovan, M.; Petrini, G.;
(
(
(
Phys. 1990, 92, 943.
(
(
(
57) Amirand, C.; Maillard, D. J. Mol. Struct. 1988, 176, 181.
58) Ayers, G. P.; Pullin, A. D. E. Spectrochim. Acta 1976, 32A, 1641.
59) Schriver, A.; Silvi, B.; Maillard, D.; Perchard, J. P. J. Phys. Chem.
(
(
1
977, 81, 2095.
Genoni, F.; Venturelli, B.; Zecchina, A.; Bordiga, S.; Ricchiardi, G. Stud.
Surf. Sci. Catal. 1994, 84, 559.
(60) Barnes, A. J.; Hallam, H. E.; Scrimshaw, G. F. Trans. Faraday
Soc. 1969, 65, 3172.
(21) DISCOVER, version 95.0; MSI Technology Inc.: San Diego, CA,
October 1995.
(61) Josien, A.; Sourisseau, B. Bull. Soc. Chim. 1955, 178.
(62) West, W.; Arthur, P. J. Chem. Phys. 1937, 5, 10.
(63) West, W.; Edwards, R. T. J. Chem. Phys. 1937, 5, 14.
(64) Andrews, L.; Hunt, R. D. J. Chem. Phys. 1988, 89, 3502.
(65) Kirkwood, J. G. J. Chem. Phys. 1934, 2, 351.
(66) Johnson, G. L.; Andrews, L. J. Phys. Chem. 1983, 87, 1852.
(
(
22) Hill, J.-R.; Sauer, J. J. Phys. Chem. 1994, 98, 1238.
23) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G.
A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,
V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;