Activation conditions play a key role in the activity of zeolite CaY: NMR and product studies of Bronsted acidity

Article abstract of DOI:10.1021/jp980385w

CaY, activated under different conditions, was characterized with ¹H, ³¹P, and ¹H/²⁷A] double resonance MAS NMR. The ¹H MAS NMR spectra of CaY, calcined in an oven at 500 °C, shows resonances from H₂O (bound to Ca²⁺ and the zeolite framework), CaOH⁺, aluminum hydroxides, silanols, and Bronsted acid sites. No evidence for Lewis acidity is observed on adsorption of trimethylphosphine, and an estimate of ≈16 Bronsted acid sites per unit cell is obtained for this sample. CaY activated in an oven at higher temperatures contains less water, but all the other species are still present. In contrast, CaY activated by slow ramping of the temperature under vacuum to 500 or 600 °C shows a much lower concentration of Bronsted acid sites (<1/unit cell). Again, no evidence for Lewis acidity was observed. These NMR results have been utilized to understand the very different product distributions that are observed for reactions of 1,1- and 1,2-diarylethylenes in zeolite CaY activated in an oven (in air) and under vacuum. Samples with high concentrations of Bronsted acid sites react stoichiometrically with these sites, yielding diarylalkanes. At low concentrations, the Bronsted acid sites can act catalytically resulting in isomerization reactions.

Full text of DOI:10.1021/jp980385w

J. Phys. Chem. A 1998, 102, 5627-5638
5627
Activation Conditions Play a Key Role in the Activity of Zeolite CaY: NMR and Product
Studies of Brønsted Acidity
Hsien-Ming Kao, Clare P. Grey,* Kasi Pitchumani, P. H. Lakshminarasimhan, and
V. Ramamurthy*
Departments of Chemistry, State UniVersity of New York, Stony Brook, New York 11794-3400, and
Tulane UniVersity, New Orleans, Louisiana 70118-5698
ReceiVed: December 2, 1997; In Final Form: February 5, 1998
CaY, activated under different conditions, was characterized with 1H, 31P, and 1H/27Al double resonance MAS
1
NMR. The H MAS NMR spectra of CaY, calcined in an oven at 500 °C, shows resonances from H2O
(bound to Ca2+ and the zeolite framework), CaOH+, aluminum hydroxides, silanols, and Brønsted acid sites.
No evidence for Lewis acidity is observed on adsorption of trimethylphosphine, and an estimate of ≈16
Brønsted acid sites per unit cell is obtained for this sample. CaY activated in an oven at higher temperatures
contains less water, but all the other species are still present. In contrast, CaY activated by slow ramping of
the temperature under vacuum to 500 or 600 °C shows a much lower concentration of Brønsted acid sites
(<1/unit cell). Again, no evidence for Lewis acidity was observed. These NMR results have been utilized
to understand the very different product distributions that are observed for reactions of 1,1- and
1,2-diarylethylenes in zeolite CaY activated in an oven (in air) and under vacuum. Samples with high
concentrations of Brønsted acid sites react stoichiometrically with these sites, yielding diarylalkanes. At low
concentrations, the Brønsted acid sites can act catalytically resulting in isomerization reactions.
Introduction
of a small amount of water to the dehydrated material has been
observed to increase the activity, while a decrease in activity
was observed with increased calcination temperatures. These
observations could be rationalized by the formation, on adding
water, of a number of Brønsted acid sites and MOH+, where
M represents an alkaline-earth divalent cation, followed by the
subsequent recombination of the MOH+ and acid sites to release
water, on heating. Infrared studies of zeolite Y containing Ca2+
and Mg2+ and a small amount of adsorbed water have been
reported.2-4 An absorption band was observed at around 3600
cm-1 in dehydrated CaY, which increased significantly in
intensity after readsorption of water, and was attributed to
CaOH+ groups.3,4 X-ray diffraction has been performed to
determine the distribution of Ca2+ cations on the different cation
positions in a partially dehydrated sample of zeolite CaNaY
(Si/Al ) 2.37).9,10 Ca2+ was located on site I′ and II, and H2O
on site II′. Recently, Hunger et al. have used 1H MAS (magic
angle spinning) NMR to study the mechanisms of the formation
of hydroxyl groups in MgNaY and CaNaY.5 On the basis of
an empirical relation, obtained by Yesinowski et al.,11 for the
isotropic chemical shift of hydroxyl groups, versus the O-H‚‚‚O
bond lengths (where H‚‚‚O respresents the hydrogen bond) they
concluded that the resonances at 0.5 and 2.8 ppm observed in
One of our groups has been investigating a number of
reactions of unsaturated organic molecules inside the pores of
zeolites such as X and Y.1 In several of these studies involving
divalent exchanged forms of the zeolites, the reactivities of the
molecules were found to be very sensitive to the method of
activation. For example, aryl olefins sorbed in activated CaY
zeolites were observed to react to form the corresponding aryl
alkanes, when CaY was activated in an oven. Very little
reaction was observed for reactions using zeolites activated
under a vacuum. These initial observations prompted a more
detailed study of this phenomenon and the characterization of
CaY activated under different conditions.
Divalent cation-exchanged zeolites have been previously
studied with a variety of spectroscopy methods.2-5 In this study,
we extend the previous work on this, and similar zeolites, by
carefully correlating the species that are present (e.g., CaOH+,
Brønsted acid sites) with the catalytic activity of the system. In
particular, the dramatic differences between oven and vacuum
activation are explored. 1,1- and 1,2-diarylethylenes are used
as chemical probes of CaY activity, and variable-temperature
1H MAS and 1H/27Al and 31P/27Al TRAPDOR NMR6,7 are
utilized to characterize the active sites (e.g., the Brønsted and
Lewis acid sites) present in CaY. 1H/27Al and 31P/27Al
TRAPDOR NMR methods are used to probe the H-Al and
P-Al distances and are useful assigment tools.6,7 Trimeth-
ylphosphine was sorbed on CaY, and 31P/27Al TRAPDOR NMR
was used to probe Lewis acidity.
1
the H MAS NMR spectrum of a CaNaY sample were due to
CaOH+ groups in the supercages and sodalite cages, respec-
tively. In this report, we establish that CaY activated under air
and under vacuum possess drastically different numbers of acidic
sites and very different species; this is related to the catalytic
properties.
Dehydrated divalent cation-exchanged zeolites are known to
act as catalysts for cracking and alkylation reactions, due to
the acidic species formed by the dissociation of water.8 Addition
Experimental Section
Preparation of CaY. NaY (Aldrich, 10 gm) was added to
200 mL of a 10% Ca(NO3)2 solution, and the slurry was stirred
at 80 °C overnight. The slurry was filtered and the filtrate
* To whom correspondence should be addressed: C.P.G., State Univer-
sity of New York; V.R., Tulane University.
S1089-5639(98)00385-5 CCC: $15.00 © 1998 American Chemical Society
Published on Web 04/22/1998

5628 J. Phys. Chem. A, Vol. 102, No. 28, 1998
Kao et al.
TABLE 1: Different Samples Studied by NMR with the
washed thoroughly with at least 2 L of distilled water. The
above exchange procedure was repeated at least four times. The
filtrate obtained after the wash was dried at room temperature
in air. ICP analysis indicated that the zeolite thus obtained has
Activation Conditions, Loading Levels, and Nomenclature
Used to Label the Samples Listed for the Different Samples
name
activation conditions
loading level
a composition corresponding to Si138.7Al53.3Na7.5Ca23.3O384
.
bare CaY
none
none
none
CaY-500-oven
CaY-600-ovena
500 °C in oven, 110 °C
in a vacuum
600 °C in oven, 110 °C
in a vacuum
500 °C in a vacuum
200 °C in a vacuum
600 °C in a vacuum
500 °C in oven, 110 °C
in a vacuum
600 °C in oven, 110 °C
in a vacuum
Preparation of D2O-Exchanged CaY. A known amount
of CaY was stirred in D2O at 60 °C for 6 h. The sample was
filtered and once again stirred in D2O at 60 °C. This process
was repeated three times, and the final sample was stored under
nitrogen. Activation was done under a nitrogen atmosphere.
Preparation of 1,1-Diarylethylenes (1, 4, and 7). 1,1-
Diphenylethylene (1a) was purchased from Aldrich and used
without further purification. All other olefins were prepared
from the corresponding ketones by a Grignard reaction followed
by a dehydration. The standard procedure adopted is described
below: A 1 equiv amount of ketone was taken in dry diethyl
ether solution under nitrogen atmosphere. To this, 1.5 equiv
of methylmagnesium bromide (3 M solution in ether from
Aldrich) was added, and the solution was stirred for another 1
h. The ether solution was then quenched with 0.1% aqueous
HCl. The organic portion was extracted into ether and
concentrated. The alcohol thus obtained was dehydrated by
refluxing in 30 mL of benzene with a catalytic amount of
p-toluenesulfonic acid. The product was purified by column
chromatography by using petroleum ether/dichloromethane as
the eluant. Required Grignard reagents for 1,1-diarylethylenes
4 and 7 were prepared from the corresponding ethyl bromide
or isopropyl bromide, respectively.
none
CaY-500-vac
none
none
none
33 TMP/uc
(unit cell)
24 TMP/uc
CaY-200-vacb
CaY-600-vac
TMP/CaY-500-oven
TMP/CaY-600-oven
TMP/CaY-500-vac
TMP/CaY-600-vac
TMP/HY-500-oven
500 °C in a vacuum
600 °C in a vacuum
500 °C in oven, 110 °C
in a vacuum
22 TMP/uc
19 TMP/uc
30 TMP/uc
a Sample was directly put into an oven held at 600 °C for 16 h.
b Sample was calcined in a vacuum at 200 °C for 6 h.
Activation of Zeolites for the Product Studies. Two
methods of activation were adopted: (A) in an oven under air;
(B) on a vacuum line at a reduced pressure (5 × 10-4 Torr).
Method A: A known amount of zeolite was placed in a
porcelain crucible in a preheated oven (100 to 700 °C) and
activated under aerated conditions for approximately 10 hours.
Method B: A known amount of zeolite was placed in a quartz
tube which was then attached to a vacuum line via Teflon
fittings. The sample was degassed at room temperature for 3 h
under reduced pressure (5 × 10-4 Torr) and then slowly heated
(15 °C/5 min) under reduced pressure to the desired temperature
(250-400 °C) and kept at that temperature for 10 h.
Inclusion of Olefins in Zeolites. The activated zeolite was
cooled to room temperature and added to a hexane solution of
the olefin (typically 20 mg of olefin for 300 mg of zeolite;
hexane, 5 mL) and stirred for 3 h. The products were extracted
from the zeolite by stirring the zeolite overnight with either
dichloromethane or tetrahydrofuran. Products were analyzed
by GC (Hewlett-Packard 5890; SE-30 capillary column). For
identification purposes, the products were isolated by flash
chromatography and characterized by NMR and GCMS.
Sample Preparation for NMR Experiments. Zeolite
samples were dehydrated either in an oven or in a vacuum by
slowly ramping the temperature to 500 or 600 °C over a period
of 16 h. The samples were then held at this temperature for
another 16 h. Some materials, prepared in the oven, were
subsequently dehydrated at 110 °C overnight under vacuum to
remove any residual weakly bound or physisorbed water.
Trimethylphosphine (TMP, 99%, Alfa) was introduced into
some samples at room temperature, over a vacuum line, to
investigate the Brønsted and Lewis acidity. Details of the
different sample preparations are given in Table 1. A sample
of HY (Aldrich) (denoted TMP/HY-500-oven) was also acti-
vated and loaded with TMP, for comparison with the CaY
studies. All samples were packed in a glovebox, under dry
nitrogen, into MAS rotors for the NMR experiments.
1
Figure 1. Pulse sequences used in the (a) H/27Al and (b) 31P/27Al
TRAPDOR NMR experiments. The π/2 and π pulses are synchronized
with the rotor period, in both (a) and (b). Intensities are measured from
experiments performed with (I) and without 27Al irradiation (I0), the
latter providing the control experiment.
NMR Experiments. NMR spectra were acquired on a
Chemagnetics CMX-360 spectrometer with a triple-tuned Che-
magnetics probe, tuned for 1H (360 MHz), 31P (142 MHz), and
27Al (93 MHz). 1H and 31P π/2 pulse lengths of typically 5
and 6.5 µs and repetition times of 2 s were used. Chemical
shifts for 31P spectra are quoted relative to 85% H3PO4 as an
external reference. An 27Al rf amplitude (ν1) of 55 kHz was
used for the 1H/27Al and 31P/27Al TRAPDOR NMR experiments
shown in Figure 1a,b. Refocusing at the rotor echo, of the I )
1/2 nuclei (1H or 31P) that are coupled to quadrupolar nuclei S
(27Al), is prevented by the combination of magic angle spinning
(MAS) and continuous irradiation of the S nuclei. This results
in a loss of I-spin intensity at the spin-echo. More detailed
theoretical descriptions of the TRAPDOR NMR experiment and
some applications can be found in the refs 6, 7, and 12.

Activity of Zeolite CaY
J. Phys. Chem. A, Vol. 102, No. 28, 1998 5629
TABLE 2: Observations upon Inclusion of 1,1-Diphenylethylene within M2+Y Zeolites: Variation of Activation Conditions,
Loading Level and Nature of Cation
product (%)b,h
diphenylethylene
remaining (%)
diphenylethane
benzophenone
conditiona
(2)
(3)
colorc
white
(a) Variation of Activation Conditions: CaYd
CaY-100-oven
CaY-150-oven
CaY-200-oven
CaY-300-oven
CaY-400-oven
CaY-500-oven
CaY-600-oven
CaY-750-oven
CaY-400-vacuume
CaY-450-vacuume
CaY-550-vacuume
98
95
69
15
15
5
2
1
1
1
1
4
30
74
74
85
32
4
light green
light green
green
green
green
60
96
92
95
98
green
light green
very light green
very light green
very light green
7
4
1
(b) Variation of Loading Level: CaY-500-oven
1 molecule/10 cages
1 molecule/4 cages
1 molecule/1 cage
1 molecule/0.5 cage
1
78
78
74
69
1
1
1
1
green
green
green
green
1
5
15
(a) Variation of Cation: M2+Y-500-Oven
MgY
60
40
99
99
80
38
57
1
<1
SrY
2
BaY
MgX; CaX; SrX; BaXf
CaY/solidg
20
green
a The data presented here, except where it is stated to be “solid”, correspond to a condition where the mixing was done in a hexane slurry. b The
yield corresponds to the amount found in the reacted material. Yield is based on GC analysis in the presence of an internal standard. Isolated yield
varied between 80 and 95%. c The color of the slurry after addition of diphenylethylene to the activated zeolite. d The amount of diphenylethylene
was maintained at one molecule per two supercages. e The sample was activated on a vacuum line at 10-4 Torr. f None of the M2+X zeolites
showed behavior similar to that of M2+Y. g The solid zeolite and diphenylethylene were ground on a mortor and left in a vial. The green color
appeared immediately upon mixing. h The yields of dimeric alkene and an indenyl product are not provided.
TABLE 3: Obervations upon Inclusion of 1,1-Diarylethylenes within M2+Y Zeolites: Generality of the Phenomenon
product (%)b
olefin
remaining (%)
diphenylethane benzophenone
diffuse reflectance
max (nm)
(2)
(3/6)
colorc
p,p’-dimethoxy DPE (1b)
p-methoxy DPE (1c)
p,p’-dimethyl (1d)
78
39
68
21
57
24
1
4
7
bright orange
cream
bright yellow
486 (strong), 680 (weak)
452 (strong), 653 (weak)
457 (strong),
642 (very weak)
p,p’-dichloro (1e)
95
95
48
77
98
95
2
1
20
5
2
4
3
4
greenish yellow 462 (strong), 630 (weak)
p,p’-difluoro (1f)
cream
436 (weak), 610 (weak)
1,1-diphenyl-1-propene (4a)
32 (6a)
18 (6b)
orange
444
496
454
498
1,1-bis(p-methoxyphenyl-1-propene (4b)
1,1-diphenyl-2-methyl 1-propene (7a)
1,1-bis(p-methoxyphenyl)-2-methyl-1-propene (7b)
bright orange
orange
bright orange
Diffuse Reflectance Spectra. DR spectra were recorded by
packing the sample in a 1 mm quartz cell. The background
correction was done by recording a spectrum of BaSO4 in the
same cell. The recorded spectrum for the sample was converted
into Kubelka Munk by the program supplied with the instrument
(Shimadzu model 2101 PC).
at temperatures between 400 and 550 °C. The majority of the
experiments were carried out with the sample CaY-500-oven.
When CaY-500-oven was added to a hexane solution of 1,1-
diphenylethylene, the zeolite-hexane slurry turned yellow and
then green and remained green for several days. This phenom-
enon was general to all the diarylethylenes 1b-f, and a variety
of brightly colored zeolitic samples were produced. The
products were extracted from these samples with either dichlo-
romethane or tetrahydrofuran and were identified by their NMR
and mass spectral characteristics. Diarylethanes (2a-f) were
the major products isolated, along with small amounts of the
corresponding diaryl ketones (Tables 2 and 3; Scheme 1), and
the zeolite was left with a blue color which persisted for at least
10 days. Extraction of the zeolite upon acid treatment results
in two additional products: adimeric alkene and an indenyl
product. Substituents on the aryl groups resulted in reduced
conversions, the smallest conversions being observed for the
difluoro (1f) and dichloro (1e) compounds.
Results
Product Studies. Two sets of olefins were used as probes
to examine the reactivity of CaY. The first set of olefins are
the 1,1-diarylethylenes (1a-f), and the second set are the 1,2-
diarylethylenes (9a-d and 11a-d). The substituents are present
in the para-position of the aryl ring in both cases. CaY was
activated by using two different methods. In the first, designated
as CaY-temp-oven, the zeolite was activated in air in an oven
at a specified temperature (between 100 and 750 °C). For
example, CaY-500-oven refers to a sample that was activated
at 500 °C in an oven. In the second method, designated as
CaY-temp-vacuum, the zeolite was activated on a vacuum line
The reactivity of diarylethylenes in CaY was very sensitive
to the method used to activate CaY (Table 2). The reaction

5630 J. Phys. Chem. A, Vol. 102, No. 28, 1998
Kao et al.
SCHEME 1: Products upon Inclusion of Diarylethylenes
within Activated CaY
Figure 2. Diffuse reflectance spectra of diphenylethylene and 1,1-
diphenyl ethane-1-ol included within activated CaY-500-oven: diphe-
nylethanol (continuous line); diphenylethylene (dashed line); diphe-
nylethylene after washing with methylene chloride (broken line).
Figure 3. Diffuse reflectance spectra of 1,1-bis(4-methoxyphenyl)-
ethylene and 1,1-bis(4-methoxyphenyl)ethan-1-ol in CaY-500-oven:
2,2-bis(p-methoxyphenyl)ethanol (continuous line); 1,1-bis(4-methox-
yphenyl)ethylene (dashed line); 1,1-bis(4-methoxyphenyl)ethylene in
CaY after washing with methylene chloride (broken line).
was nearly quantitative toward reduction of 1a when the sample
CaY-500-oven was used (Table 2a). Surprisingly, reactions
performed with CaY-750-oven and CaY-400-vacuum showed
very little activity: When these zeolites were added to 1a, no
bright coloration was observed; nor was a significant amount
of the reduced product 2 formed. CaY was least active when
it was activated in an air oven below 150 °C or when it was
activated on a vacuum line above 400 °C. Low activity of the
former sample is most likely due to the presence of residual
water which prohibits the entry of an olefin into the zeolite.
The effect of the loading level of 1a in CaY-500-oven on the
reactivity was also explored (Table 2b). Little difference in
the percent of diarylethane formation was observed at low
loading levels. A slight reduction in product formation was
observed when the loading level was increased from 1 molecule
in 4 supercages to 1 molecule per cage.
500-oven (Scheme 1). Again, consistent with the observations
made for the 1,1-diarylethylenes, addition of CaY-450-vacuum
to trans-1,2-diarylethylenes 9a-d, resulted in no observable
reaction. Surprisingly, cis-1,2-diarylethylenes 11a-d were
quantitatively isomerized to the trans isomer 9a-d, in CaY-
450-vacuum, without any reduction taking place.
Characterization of Reactive Intermediates by Diffuse
Reflectance Spectroscopy. Diffuse reflectance spectra were
acquired in order to study the cause of the deep colors formed
on addition of the diarylethylenes. The diffuse reflectance
spectra displayed in Figures 2 and 3, for 1,1-diphenylethylene
(1a) and 1,1-bis(4-methoxyphenyl)ethylene (1b), respectively,
consist of two distinct maxima (one below 500 nm and the other
above 600 nm). Following extraction of the products from the
zeolite with either diethyl ether or tetrahydrofuran, which
changes the color of the zeolite from green to blue, the spectra
consist mainly of the absorption above 600 nm. This suggested
that two independent species are responsible for the two
absorptions. One possible assignment of the intermediate
responsible for the shorter wavelength absorption (428 nm for
1,1-diphenylethylene and 486 nm for 1,1-bis(4-methoxyphenyl)-
The reactivity of 1a in a series of divalent cation exchanged
X and Y zeolites was also explored (Table 2c), and the activity
was observed to decrease in the order Ca > Sr > Mg > Ba for
the Y series. Negligible activity was observed in the X series
in comparison to the Y series, BaY showing behavior similar
to those of X zeolites.
Similar to reactions of 1, olefins 9a-d and 11a-d were
reduced to the corresponding 1,2-diarylethanes (10) in CaY-

Activity of Zeolite CaY
J. Phys. Chem. A, Vol. 102, No. 28, 1998 5631
Figure 4. Time-dependent diffuse reflectance spectra of diphenyleth-
ylene in CaY-500-oven. Note the development of the peak at 618 nm
with time: (continuous line) within 15 min of addition; (dashed line)
20 h after addition; (broken line) 40 h after addition.
Figure 5. Diffuse reflectance spectra of 1,1-bis(4-methoxyphenyl)-
prop-1-ene (4b) (continuous line) and 2-methyl-1,1-bis(4-methoxyphe-
nyl)prop-1-ene (7b) (dashed line) in CaY-500-oven.
Isotope Studies of the Reaction Mechanism. The major
product upon inclusion of 1,1-diarylethylenes (1a-f) and 1,2-
diarylethylenes (9a-d and 11a-d) within CaY results from the
addition of two hydrogen atoms to the π-bond. One possibility
for the first step in the reduction process involves the addition
of a proton to the olefin (Scheme 2). This suggestion is
consistent with the diffuse reflectance measurements, where the
carbocation was observed as a stable intermediate. To test this,
and other possible reaction mechanisms, reactions were per-
formed with CaY that had previously been washed with D2O
and CaY(D2O) to remove the H2O and to deuterate any water
or hydroxyl groups that remain following activation. Sorption
and reaction studies were performed in deuterated solvents, and
the products were analyzed with mass spectroscopy to determine
the level and positions of any deuterium incorporation in the
products. Results from these studies were contrasted with those
obtained with CaY(H2O) in deuterated solvents (C6D12 and CD2-
Cl2) and are shown in Scheme 3.
SCHEME 2: Brønsted Acid-Site-Mediated Reduction of
Diarylethylenes
Figure 6a shows the mass spectral fragmentation pattern for
diphenylethane obtained in a deuterated solvent after extraction,
when nondeuterated CaY(H2O) was used. The mass ion at m/e
182 and the base peak at m/e 167, due to CH3 loss, are clearly
seen, consistent with presence of 1,1-diphenylethane with a “D”
in carbon-1. In contrast, when CaY(D2O) was used, a mass
ion at m/e 183, and a base peak at m/e 168 (Figure 6b) were
observed, which result from 1,1-diphenylethane in which the
carbon-2 was substituted with a “D” instead of an “H”. This is
consistent with the suggestion that the initial protonation occurs
at the carbon-2 of diphenylethylene and that the proton comes
from acid sites present in CaY via the included H2O or D2O.
Characterization of CaY by NMR: Quantification of
Acidic Sites. CaY-500-oVen. Figure 7 shows the variable-
ethylene), based on literature reports,13 is the carbocation 12,
derived via the protonation of the parent olefin (Scheme 2).
This was confirmed by generating the same carbocations (12)
independently from the corresponding carbinols 13 (Scheme 2).
The spectra from 1,1-diphenyl ethan-1-ol and 1,1-bis(4-meth-
oxyphenyl)ethan-1-ol sorbed in CaY are shown in Figures 2
and 3, respectively, and are nearly identical at short wavelengths.
The time-dependent diffuse reflectance spectra show changes
in the relative intensities of the two absorptions as a function
of time (Figure 4). The absorption assigned to the monomer-
cation 12 dominates the time-dependent diffuse reflectance
spectra of diphenylethylene in CaY-500-oven (Figure 4) at early
times, but with time, the second absorption at 618 nm grows in
at the expense of the monomer-cation peak. The identity of
the species responsible for this absorption is discussed in a later
section.
The diffuse reflectance spectra of the 1,2-diarylethylenes (9
and 11) in CaY showed absorptions in the region 300-370 nm,
and no longer wavelength absorptions were present. The
absorption in the region 300-370 nm is attributed to a benzyl
carbocation derived by the addition of a proton to 9 and 11.14
Inclusion of 4 and 7 (Scheme 1) within CaY-500-oven gave a
bright orange color (Table 3). The diffuse reflectance spectrum
of this yellow solid consisted of a single absorption as shown
in Figure 5.
1
temperature H MAS NMR of a CaY sample (CaY-500-oven)
calcined in an oven at 500 °C. Only one major resonance at
3.7 ppm, with small spinning sidebands and a shoulder at 4.5
ppm, is observed at room temperature. The former observation
suggests that this resonance results from a number of species
in rapid exchange with each other. Very low intensity spinning
sideband manifolds, which spread over more than 100 ppm,
are also visible when the spectrum is enlarged. These are
assigned to water rigidly held to the zeolite lattice. On cooling
of the sample, more resonances become visible as the exchange
process between water and the other species starts to freeze out.
At -150 °C, three major resonances at 3.7, 2.4, and 1.2 ppm

5632 J. Phys. Chem. A, Vol. 102, No. 28, 1998
Kao et al.
Figure 6. Mass spectra of the four H/D isomers of diphenylethane. See Scheme 3 for details.
SCHEME 3: Possible Sources of Hydrogens during the Conversion of Diarylethylenes to Diarylethanes
1
doublet” (due to two isolated, strongly coupled H spins) are
also observed; this is consistent with the sideband pattern
expected for two protons in a rigidly bound water molecule,
the sidebands resulting from the proton-proton dipolar coupling
between the two protons. This resonance, which has an isotropic
chemical shift close to that of the Brønsted acid sites, is
tentatively assigned to water bound to a calcium cation, i.e.,
2+
Ca(OH2)n (n ) 1).
1H/27Al TRAPDOR NMR experiments have been performed
on this sample as shown in Figure 8. The intensities of the
resonances of proton species that are nearby aluminum atoms
are expected to decrease in the spectrum acquired with 27Al
irradiation. This occurs because the dephasing during the
1
evolution time of the spin-echo experiment, due to H-27Al
dipolar coupling, is no longer refocused on applying 27Al
irradiation.12 Thus, the only proton species observed in the
difference experiment are those that are nearby to aluminum
atoms. As expected, the Brønsted acid sites (3.7 ppm) and Al-
1
OH groups (2.4 ppm) show large H/27Al TRAPDOR effects
in the difference spectrum (Figure 8c). Two resonances at 2.1
and 1.2 ppm, with weak and significantly more intense spinning
sidebands, respectively, are observed in the spectrum obtained
with 27Al on-resonance irradiation (Figure 8b), implying that
that these two species are not in close proximity to aluminum
atoms. Two different CaOH+ groups that resonate at 0.5 and
2.8 ppm have been reported previously and have been assigned
to groups located in the supercages and sodalite cages,
respectively.5 Unfortunately, these resonances are very close
in chemical shift to those of Al-OH and Si-OH species and
it is difficult to assign these groups unambiguously. Neither
of the species that resonate at 2.1 ppm or at 1.2 ppm can be
1
Figure 7. Variable-temperature H MAS NMR spectra of CaY-500-
oven collected at a spinning speed of 3 kHz.
and a shoulder at approximately 5 ppm are observed. The
resonances at 3.7, 5.0 and 2.4 ppm are assigned to the Brønsted
acid sites (3.7 and 5.0 ppm) and Al-OH groups (2.4 ppm).
Spinning sidebands with an envelope resembling a “Pake

Activity of Zeolite CaY
J. Phys. Chem. A, Vol. 102, No. 28, 1998 5633
Figure 8. 1H/27Al TRAPDOR spectra of CaY-500-oven at -150 °C
obtained (a) without and (b) with on-resonance 27Al irradiation during
τ (τ ) 333 µs, spinning speed ) 3 kHz). The difference spectrum is
shown in (c). The intensity of (c) has been scaled by a factor of 2 with
respect to (a) and (b).
Figure 9. Variable-temperature31P {1H} MAS NMR spectra of TMP/
CaY-500-oven collected at a spinning speed of 5260 Hz.
assigned to Al-OH groups, however, since they are observed
in the double resonance experiment (Figure 8b), and the
TRAPDOR fractions (1 - I/I0), where I0 and I represent the 1H
echo intensities obtained without and with 27Al irradiation,
respectively, for these species are expected to be close to 1.0,
under conditions used in our experiments.6 On the basis of this,
and the chemical shift, we tentatively assign the resonance at
2.1 ppm to a CaOH+ and the resonance at 1.2 ppm to a silanol
1
group. A significant H/27Al TRAPDOR effect was seen for
the Pake doublet (Figure 8c) indicating that the water molecules
are in close proximity to aluminum atoms.
TMP/CaY-500-oVen. The variable-temperature 31P proton-
decoupled MAS NMR spectra of the sample TMP/CaY-500-
oven are shown in Figure 9. Two broad resonances at -4 and
-61 ppm were observed at room temperature. As the temper-
ature decreases, the resonance at -4 ppm shifts to higher
frequency and grows in intensity. Small spinning sidebands
are observed at -100 °C and below. In contrast, the resonance
at -61 ppm initially shifts slightly and sharpens (-20 °C). At
lower temperatures, other smaller resonances become visible,
and at -150 °C, the major resonance has shifted to -66 ppm
and has two shoulders at -56 and -61 ppm. The resonance at
-2 ppm has been assigned to a TMPH+ species, resulting from
the reaction of TMP and the Brønsted acid sites, and the
resonance at -66.0 ppm was assigned to liquidlike TMP
species.15,16 TMP bound to an aluminum Lewis acid sites is
expected to resonate in the range of -46 to -50 ppm,7 and the
absence of a resonance in this range suggests that there are no
accessible Lewis acid sites for TMP. 31P/27Al TRAPDOR NMR
experiments were performed, to confirm that the additional
resonances are not in fact due to TMP binding to aluminum
Lewis acid sites (Figure 10). As seen in the difference spectrum
(Figure 10c), only the resonance at -2 ppm assigned to TMPH+
shows a significant loss of intensity on 27Al irradiation.
Figure 10. 31P/27Al TRAPDOR NMR spectra of TMP/CaY-500-oven
obtained (a) without (b) with on-resonance 27Al irradiation. The
difference spectrum is shown in (c). (spinning speed ) 5040 Hz; 27Al
rf field ) 55 kHz; 27Al irradiation time ) 398 ms).
1
CaY-600-oVen. Figure 11 shows the room-temperature H/
27Al TRAPDOR spectra of a sample CaY-600-oven that was
calcined in oven at 600 °C directly without any ramping period.
The concentration of water in this sample has diminished
considerably, as seen by the existence of significant spinning
sideband manifolds, even at room temperature. Three major
resonances at 3.4, 2.8, and 1.2 ppm were observed in the control
experiment (Figure 11a). Three resonances were seen at 3.4,
2.4, and 0.8 ppm, with a shoulder at 3.7 ppm, in the difference
spectrum (Figure 11c) indicating that these resonances arise from
species near aluminum atoms. Both the resonances at 3.4 (with

5634 J. Phys. Chem. A, Vol. 102, No. 28, 1998
Kao et al.
Figure 12. 1H/27Al TRAPDOR NMR spectra of CaY-500-vac obtained
(a) without (b) with 27Al on resonance irradiation. The difference
spectrum is shown in (c) (spinning speed ) 3 kHz; 27Al rf field ) 55
kHz; 27Al irradiation time ) 333 µs).
Figure 11. 1H/27Al TRAPDOR NMR spectra of CaY-600-oven
obtained (a) without (b) with 27Al on-resonance irradiation. The
difference spectrum is shown in (c) (spinning speed ) 3 kHz; 27Al rf
field ) 55 kHz; 27Al irradiation time ) 333 ms).
detectable concentration of TMPH+ is observed. On the basis
of our earlier study of TMP adsorption in activated NaY zeolites,
we estimate that we are sensitive to levels of ≈1 Brønsted acid
site per unit cell with this method,17 and hence, less than these
detectable levels are present in these samples. Again, no
resonances that can be assigned to TMP bound to the aluminum
Lewis acid sites were detected. Figure 12 shows the room-
temperature 1H/27Al TRAPDOR NMR spectra of CaY-500-vac.
Only a negligible concentration of Brønsted acid sites (at around
4 ppm) was observed. The resonance at 0.9 ppm dominates
the spectra, the resonance at around 0.0 ppm being caused by
the probe head background. Similar spectra were observed for
CaY-600-vac except that even lower concentrations of Brønsted
acid sites were observed. The resonance at 0.9 ppm does not
show a significant TRAPDOR effect on 27Al irradiation (Figure
12b). Instead, a broad resonance in the range of 0.7-2.6 ppm
was observed in the difference spectrum of 1H/27Al TRAPDOR
NMR experiment (Figure 12c). Since the observed TRAPDOR
effect for the resonance at 0.9 ppm is much less than that
expected for Al-OH groups, this suggests that this resonance
is due to CaOH+ groups possibly in the supercages.
a shoulder at 3.7 ppm) and 2.4 ppm have large spinning sideband
manifolds which are similar to those expected for Al-OH
groups, and these resonances are assigned to Brønsted acid
protons and Altel-OH, respectively, based on our earlier studies
of dehydroxylated HY.6 The resonance due to Altel-OH is
clearly obscured in the control experiment, by the larger
resonance at 2.8 ppm. The resonance at 0.8 ppm is assigned
to Aloct-OH groups, again on the basis of previous studies.6
The spectrum obtained with 27Al irradiation (Figure 11b) shows
two major resonances at 2.8 and 1.2 ppm. The resonance at
2.8 ppm is associated with larger spinning sideband manifolds
than the resonance at 1.2 ppm; this is more clearly visible when
the spectrum is enlarged. This observation is similar to that
made for the resonances at 2.1 and 1.2 ppm in the spectrum of
CaY-500-oven acquired under similar conditions (Figure 8b),
and the resonance at 2.8 ppm is tentatively assigned to a CaOH+
group. Again, the resonance at 1.2 ppm is due to a silanol
group. Interestingly, the resonance at 3.4 ppm comprises
broader and sharper overlapping resonances, and the sideband
manifolds appear to be coming from the broader component.
The broader and sharper components are assigned to isolated
Brønsted acid sites and Brønsted acid sites that are undergoing
an exchange process with water molecules in the sample,
respectively.
27Al MAS NMR of the dehydrated samples shows a small
resonance at 80 ppm that is assigned to the tetrahedral aluminum
oxide anion in the sodalite cage18 and is approximately 3-5%
of the intensity of the resonance from the framework aluminum
atoms.
CaY-200-Vac. The 1H/27Al TRAPDOR NMR spectra of CaY-
200-vac are shown in Figure 13. Two major resonances were
observed at 2.1 and 3.7 ppm with shoulders at 5.0, 1.2, and 0.7
ppm. As described above, the resonances at 2.1, 1.2, and 3.7
ppm are tentatively assigned to CaOH+ groups in the sodalite
cages, silanol groups, and Brønsted acid sites, respectively. As
shown in Figure 13b, resonances at 5.5, 2.1, and 1.2 ppm with
a shoulder at 0.7 ppm were observed when 27Al irradiation was
applied. Resonances at 3.7, 2.4, and 0.8 ppm show significant
TRAPDOR effects (Figure 13c). The latter two species were
assigned to extraframework Al-OH groups. Therefore, the
resonance at 0.7 ppm can be assigned to a CaOH+ group in the
supercages, consistent with the observation for the CaY-500
TMP/HY-500-oVen. To compare the effect of oven versus
vacuum activation, HY was also activated at 500 °C. The 31P
spectrum shows an intense resonance at -2 ppm that is assigned
to TMPH+. This resonance is now considerably more intense
than the resonances in the range of -56 to -66 ppm, consistent
with the increased number of Brønsted acid sites in this sample.
Again, there are no resonances at around -46 ppm from TMP
bound to the aluminum Lewis acid sites.
TMP/CaY-500-Vac. The resonance at -60 ppm, resulting
from physisorbed TMP, dominates in the 31P spectra of TMP/
CaY-500-vac and TMP/CaY-600-vac (not shown), and no

Activity of Zeolite CaY
J. Phys. Chem. A, Vol. 102, No. 28, 1998 5635
less in the supercage (site II) in comparison to the sodalite cage
(site I′) decreasing for higher water. For example, the popula-
tions per unit cell vary from 10.6-6.6 at site II and 12.0-17.3
at site I′ for samples with 12.8-26.5 water molecules per unit
cell.9,10 Thus, in the oven-activated samples that contain
considerable amounts of water, the concentrations of CaOH+
are expected to be extremely low, and the intensity due to the
CaOH+ groups in sodalite cages will be larger than that for the
CaOH+ groups in supercages. This is consistent with the
experimental observations, where a resonance at 0.5-0.9 ppm
from the CaOH+ in the supercages was not observed in the 1H
MAS NMR of these samples. Since the intensity from this
species is expected to be extremely small, the resonance may
be obscured by the Aloct-OH species at 0.8 ppm. The CaOH+
supercage resonance is observed, however, in samples activated
in a vacuum where the populations of site II calcium cations
are known to be larger. A resonance was also seen (at 0.7 ppm)
for oven activation at 200 °C, which was assigned to the CaOH+
supercage groups. Possibly, a larger fraction of the calcium
ions is present as CaOH+ for activation at these lower
temperatures.
1
A significant H/27Al TRAPDOR effect was seen for the
2+
water molecules (Figure 8) assigned to Ca(OH2)n
. This
suggests that the water is also hydrogen bonded to oxygen atoms
that are directly bound to aluminum atoms. This is possible in
the sodalite cage, where water molecules can be simultaneously
coordinated to calcium cations in site I′ and to the zeolitic
framework occupying sites such as site II′. Indeed, calcium
occupation of site I′ has been suggested to occur only when a
small amount of residual water holds the cations in these sites
(as opposed to site I).19 The possibility of the coordination
between water and Lewis acid sites (in the supercages) was
excluded because significant concentrations of Lewis acid sites
were not detected when TMP was used as a probe molecule
(see below). However, it is possible that some of the TRAP-
DOR effect is caused by water molecules coordinated to both
Ca2+ and any extraframework tetrahedral aluminum oxide
species present in the sodalite cages. These AlO4 species have
been observed both crystallographically and by 27Al MAS NMR
in CaA.18 27Al MAS NMR of the CaY samples showed a
resonance at 80 ppm that was assigned to the tetrahedral
aluminum oxide in the sodalite cage. Although this resonance
accounts for approximately 3-5% of the total aluminum
intensity, the quadrupole coupling constant of species giving
rise to the resonance is very small, in comparison to the
framework aluminum nuclei, and hence the relative intensities
of the two species need to be scaled appropriately,20 resulting
in a very small concentration of AlO4 species in comparison to
the aluminum in framework (less than 1 per unit cell). This is
not sufficient to account for all the observed TRAPDOR effect.
Figure 13. 1H/27Al TRAPDOR NMR spectra of CaY-200-vac obtained
(a) without (b) with 27Al on-resonance irradiation. The difference
spectrum is shown in (c) (spinning speed ) 3 kHz; 27Al rf field ) 55
kHz; 27Al irradiation time ) 333 µs).
(or 600)-vac sample and is close to the result reported by Hunger
et al.5 No TRAPDOR effect was observed for the resonance
at 5.5 ppm indicating that it is probably due to mobile water
molecules.
Discussion
The MAS NMR results are discussed first. This is followed
by a discussion of the chemical behavior of CaY toward diaryl
olefins. The diaryl olefin reactions and reactivity of CaY are
interpreted by making use of the MAS NMR results and the
different species detected with this technique.
Partly on the basis of earlier assignment of the resonances at
0.5 and 2.8 ppm in a sample of CaY to two different CaOH+
groups located in the supercages and sodalite cages, respec-
tively,5 we assigned1H MAS NMR resonances observed at 2.1-
2.8 ppm in this study to CaOH+ ions present in the sodalite
cages. The intensity of the resonance at 2.1 ppm was also
observed to vary significantly from sample to sample, indicating
that the presence of the species resonating at 2.1 ppm is very
sensitive to the activation conditions, as expected. The shift
from 2.8 ppm (CaY-600-oven, spectrum collected at room
temperature) to 2.1 ppm (CaY-500-oven, spectrum collected at
-150 °C) of the CaOH+ resonance suggests that the isotropic
chemical shift of this group may depend on both the water
content and the temperature of the sample. The larger sideband
manifolds associated with these resonances, in comparison to
those of the silanol groups, probably result from dipolar coupling
to nearby water molecules; residual water is more likely to be
closer to the CaOH+ group than to the silanol group and may
be present as Ca(H2O)(OH+), consistent with the assignment.
The water in the Ca(H2O)(OH)+ species is likely to be in
exchange with other Ca(OH)+(H2O)n groups in the same cage,
especially at higher water contents. This may account for the
changes in chemical shift with sample activation temperature
and the temperature of the NMR experiment. X-ray diffraction
data showed that the population of Ca2+ cations per cavity is
No evidence for any significant Lewis acidity (due to
aluminum Lewis acid sites) was detected for the samples
prepared in an oven and with the vacuum line. This is in
contrast to our earlier study of HY, where samples heated above
400 °C showed significant Lewis acidity.6 The Lewis acidity
created under these conditions results from the dehydroxylation
of the Si-O(H)-Al linkage, and it may be that, under the
conditions of temperature and vacuum required for this, the
CaOH+ and Si-O(H)-Al groups have already recombined to
release water. The small amounts of water that are released at
relatively elevated temperatures do not appear to result in a
significant amount of steaming (and Lewis acidity), although
they may help in the formation of the extraframework AlO4
species. Interestingly, the sample of HY activated in an oven

5636 J. Phys. Chem. A, Vol. 102, No. 28, 1998
Kao et al.
SCHEME 4: Brønsted Acid-Site-Mediated Geometric
does not show any Lewis acidity. This may be a consequence
of the residual water that is present, which will coordinate or
react with any Lewis acid sites that are created during the
activation process. Finally, Brønsted acid-site protons are
observed in CaY-600-oven which are rigidly held to the
framework and do not appear to be in exchange with residual
water molecules; these sites give rise to sizable 1H/27Al
TRAPDOR effects even at room temperature. These Brønsted
acid sites are unlikely to be located in the sodalite cages (where
rapid exchange with the water is expected). It is these sites
that are expected to be most reactive in the presence of the
diarylalkenes.
Isomerization of 1,2-Diarylethylenes
1H and 31P MAS NMR clearly demonstrate the presence of
high concentrations of Brønsted acid sites in CaY activated in
an oven (≈2 per supercage, for activation at 500 °C). These
presumably result from the following reaction:
MgY may be due to the smaller Mg2+ cation. Although this
should, in theory, favor the formation of MgOH+, the smaller
Mg2+ cation, located in the SII or SI′ positions, lies closer to
the plane formed by the three coordinated oxygen atoms of the
six rings than do the larger cations. Thus Mg2+ will be more
effectively screened by these oxygen atoms. This may result
in less effective polarization of any bound water molecules and,
a consequently, less Brønsted acid sites. Thus, the number of
Brønsted acid sites may be related to both cation size and cation
location. NMR studies are, however, in progress to quantify
the number of acid sites in these and a variety of other zeolites.
The proposed mechanism for the geometric isomerization of
11 within CaY-400-vacuum is shown in Scheme 44. Again the
reaction is initiated by protonation. Absence of reduction
products suggests that either the cation generated by protonation
does not survive long enough to abstract a hydride or the number
of cations is not large enough to yield an appreciable amount
of the reduction product. Given the negligible concentrations
of Brønsted acid sites identified by 1H and 31P MAS NMR, we
believe that it is the latter. Thus the final product (reduction
or isomerization), appears to be controlled by the number of
Brønsted acid sites present within CaY. When the number of
Brønsted acid sites are large with respect to the number of olefin
molecules present in a zeolite, the acid-base equilibrium favors
the permanent formation of carbocations. These carbocations
probably abstract a hydride ion from the solvent to yield the
reduction products. When the number of Brønsted acid sites
are relatively small, not enough carbocations are generated to
yield significant amounts of reduction products. However under
such conditions the acid sites can act as catalysts and effect an
isomerization of an olefin from the thermodynamically less
stable to the more stable isomer. This is similar to reactivity
observed in NaY, where the low concentrations of Brønsted
acid sites were observed to act catalytically, to cause rearrange-
ments of a number of olefins.21
Ca2+ + H2O + Si-O-Al f Ca(OH)+ + Si-O(H)-Al
where Si-O-Al represents a portion of the zeolite framework.
In contrast, samples carefully activated at 500-600 °C under
vacuum contain negligible concentrations of Brønsted acid sites
(<1 per unit cell). These quantitative conclusions are now used
to help interpret the observations made in the reaction studies
of the diarylethylenes.
The Brønsted acid sites appear to react stoichiometrically with
the diaryl olefins sorbed in the oven activated CaY. Both diffuse
reflectance spectral data (Figures 1 and 2 and Table 2) and mass-
spectroscopy studies of deuterated CaY/solvents support the
suggestion that the first step involves the protonation of the
arylalkene to form a carbocation (e.g., 12, Scheme 2). The
absorptions at 428-486 nm that were observed with diffuse
reflectance spectroscopy and that give rise to the intense colors
of these zeolites could be assigned to this carbocation. In
addition, use of a deuterated CaY lead to incorporation of
deuterium at the C-2 position. Use of deuterated solvents for
the diaryl sorption and product extraction gave 1,1-diphenyle-
thane with a “D” in carbon-1. This indicates that the second
hydrogen required to form the diarylalkane comes from the
solvent.
Differences in behavior of CaY activated under different
conditions, we believe, are the result of differences in the
number of Brønsted acidic sites present in these structures.
Furthermore, the reaction studies of 1a performed as a function
of loading level, which showed a slight reduction in product
formation for a loading level of 2 per supercage is consistent
with the numbers of Brønsted acid sites determined from NMR.
Furthermore, this drop off with loading level, and for samples
activated at different temperatures, is consistent with a stoichio-
metric reaction involving the Brønsted acid sites and not a
catalytic reaction. The lower reactivity of the fluoro- and chloro-
substituted arylalkenes is consistent with the formation of a
carbocation as the first step of the reaction mechanism, the para-
substituents on the aryl groups acting to destabilize the
carbocation. Reactivity of the diaryl in other divalent-cation-
exchanged zeolites was observed in the order CaY > SrY >
MgY > BaY. This is presumably a consequence of the
decreasing concentrations of Brønsted acid sites in these zeolites.
The oxygen atoms in these zeolites may act as basic sites,
basicity increasing for the alkaline-earth-exchanged zeolites with
size of cation. Thus, the Brønsted acid sites in BaY may be
expected to be the least acidic. However, the much smaller
number of Brønsted acid sites that are expected for BaY, due
to the larger Ba2+ cation, are likely to be a much more important
contribution to the reduced reactivity. The lower reactivity for
The reaction of the Brønsted acid site with the diaryl olefin
results in a carbocation. This carbocation acts to provide charge
neutrality with the framework. Subsequent reaction with the
solvent must involve the extraction of hydride ion from the
solvent, and a product or intermediate with a positive charge
must be generated. One possibility is that recombination with
the OH- present as CaOH+ may occur to maintain charge
neutrality in the final set of products. This suggestion requires
further experimental testing. At this stage, the source of the
hydride, and the mechanism of the second addition, are still
unclear.
At the moment, we are not certain of the structure of the
species that is responsible for the long-wavelength absorption
(>600 nm) observed in the diffuse reflectance studies. On the
basis of the earlier studies of the behavior of vinylanisole and
indene within CaY we speculate that a dimeric cationic species
might be responsible for this absorption.22 The dimeric cationic

Activity of Zeolite CaY
J. Phys. Chem. A, Vol. 102, No. 28, 1998 5637
SCHEME 5: Proposed Structures for the Persistent
Carbocation Derived from Diphenylethylene
diphenylethylene radical cation, do not correspond with the
spectrum reported in this or previous studies. Another sugges-
tion for this colored species is an olefin-acid π-complex (16,
Scheme 5).23 Since the original proposal, this has not received
much experimental support. Suggestions that the species may
be a cation derived from a dimeric or a trimeric species have
been made.25-27 Two structures 17 and 18 (Scheme 5) have
been proposed for the dimeric cation.25a,27 While these are
likely, formation of such structures would require rearrange-
ments that are thermodynamically less favorable.25a,27 We
tentatively assign the structure responsible for the blue color to
be 19, and a proposed mechanism for its formation is shown in
Scheme 6.
A key step in the reaction sequence outlined in Scheme 6 is
the conversion of carbocation 18 to the carbocation 19. This
we believe occurs through a series of acid-base equilibria as
illustrated in Scheme 6. The reaction sequence in Scheme 6
bears a close resemblance to the behavior of indene and
vinylanisole published earlier.1a,b Olefin 20, once formed via
a proton-mediated dimerization process, can be protonated by
the acidic sites present in CaY. Protonation can occur either at
the terminal carbon leading to 18 or at the middle carbon
yielding 21. Of the two processes, the former is expected to
be favored since it generates a more stable cation. However,
the latter is also likely to take place, albeit in low yields, since
this cation will also be stabilized by adjacent phenyl groups
via formation of a nonclassical phenonium ion. Given the fact
that the cation 21 will be stabilized by neighboring phenyl
groups, direct conversion of 18 to 21, via hydrogen migration,
by passing the olefin 20 is also possible. Once 21 is formed,
its transformation to 22 would be expected via a phenyl
migration. Further reactions, as outlined in Scheme 6, would
yield the carbocation 19. We suggest that the blue species
absorbing at 610 nm is the carbocation 19. Of the various
species suggested to be responsible for the 610 nm absorption
(Scheme 5), the cationic species 17, suggested by Rooney and
Hathaway,25a and 19, suggested in this report, bear close
species may result from the reaction of a monomer cation with
a neutral monomer to form a dimer cation (of unknown struc-
ture). This speculation is supported by the behavior of 1,2-
diarylethylenes (9 and 11) which do not show an absorption at
longer wavelengths. Only absorptions in the region of 300-
370 nm were observed which were attributed to benzyl carbo-
cations formed from the addition of a proton to 9 and 11.14
End capping of olefin should disfavor dimerization, especially
within a confined space. This was also supported by the
behavior of two terminal capped diarylethylenes 4 and 7 which
only a gave a yellow color in the CaY-500-oven attributable to
a monomer carbocation (Figure 4). In general these olefins gave
the reduced products in low yields (Table 3).
A deep blue or green coloration of diphenylethylene in acidic
media (in solution as well as on silica-alumina surfaces) has
been reported previously.23-28 In the past, several suggestions
have been made for the coloration (Scheme 5). One proposal
is that the color is due to a monomer radical cation (15, Scheme
5).24 Unfortunately, the absorption spectrum for 1,1-diphenyl-
ethylene radical cation in solution has not been conclusively
identified.29 However, reported spectra, assigned to the 1,1-
SCHEME 6: Proposed Mechanism for the Formation of Persistent Carbocations 17 and 19

5638 J. Phys. Chem. A, Vol. 102, No. 28, 1998
Kao et al.
Pitchumani, K.; Corbin, D. R.; Ramamurthy, V. J. Am. Chem. Soc. 1996,
118, 8152.
(2) Habgood, H. W, J. Phys. Chem. 1965, 69, 1764.
(3) Uytterhoeven, J. B.; Schoonheydt, R.; Liengme, B. V.; Hall, W.
K. J. Catal. 1969, 13, 425.
(4) Ward, J. W. J. Phys. Chem. 1968, 72, 4211.
(5) Hunger, M.; Freude D.; Pfeifer, H.; Prager, D. Chem. Phys. Lett.
1989, 63, 221.
similarity. Structure 17 results from the migration of a “methyl”
group while 19 comes from the migration of the phenyl group.
As indicated in Scheme 6, migration of a phenyl group is more
likely due to stabilization of the intermediate via phenonium
ion formation.30 Further work is underway to confirm the
suggestion.31
(6) Kao, H. M.; Grey, C. P. J. Phys. Chem. 1996, 100, 5105.
(7) Kao, H. M.; Grey, C. P. Chem. Phys. Lett. 1996, 259, 459.
(8) (a) Miale, J. N.; Chen, N. Y.; Weisz, P. B. J. Catal. 1966, 6, 278.
(b) Venuto P. B.; Hamilton, L. A.; Landis, P. S. J. Catal. 1966, 4, 81.
(9) Costenoble, M. L.; Mortier, W. J.; Uytterhoeven, J. B. J. Chem.
Soc., Faraday Trans. I, 1978, 74, 466.
(10) Costenoble, M. L.; Mortier, W. J.; Uytterhoeven, J. B. J. Chem.
Soc., Faraday Trans. I, 1978, 74, 477.
(11) Yesinowski, J. P.; Eckert, H.; Rossman, G. R. J. Am. Chem. Soc.
1988, 110, 1367.
(12) Grey, C. P.; Vega, A. J. Am. Chem. Soc. 1995, 117, 8232.
(13) (a) McClelland, R. A.; Kangasabapathy, V. M.; Steenken, S. J.
Am. Chem. Soc. 1988, 110, 6913. (b) Azarani, A.; Berinstain, A. B.;
Johnston, L. J.; Kazanis, S. J. Photochem. A: Chem. 1991, 57, 175.
(14) McClelland, R. A.; Chan, C.; Cozens, F.; Modro, A.; Steenken, S.
Angew. Chem., Int. Ed. Engl. 1991, 30, 1337.
(15) (a) Ward, J. W. In Zeolite Chemistry and Catalysis; Rabo, J. A.,
Ed.; American Chemical Society: Washington, DC, 1976; p 118. (b)
Poutsma, M. L. In Zeolite Chemistry and Catalysis; Rabo, J. A., Ed.;
American Chemical Society: Washington, DC, 1976; p 437.
(16) (a) Rothwell, W. P.; Chen, W.; Lunsford, J. H. J. Am. Chem. Soc.
1984, 106, 2452. (b) Lunsford, J. H.; Rothwell, W. P.; Shen, W. J. Am.
Chem. Soc. 1985, 107, 1540.
(17) Rao, V. J.; Perlstein, D. L.; Robbins, R. J.; Lakshminarasimhan,
P. H.; Kao, H.-M.; Grey, C. P.; Ramamurthy, V. J. Chem. Soc., Chem.
Commun. 1998, 269.
(18) (a) Corbin, D. R.; Farlee, R. D.; Stucky, G. D. Inorg. Chem. 1984,
23, 2922. (b) Pluth, J. F.; Smith, J. V. J. Am Chem. Soc. 1983, 105, 1192.
(19) Dempsey, E. J. Phys. Chem. 1969, 73, 3666.
Conclusions
Variable-temperature 1H MAS NMR spectra of CaY, calcined
in an oven, show resonances of Brønsted acid sites, Al-OH
species, and Si-OH groups. A “Pake doublet” was also
observed, with an isotropic chemical shift close to that of the
Brønsted acid sites, resulting from proton-proton dipolar
coupling between the two protons in a rigidly bound water
molecule. This resonance is assigned to water bound to a
2+
calcium cation i.e., Ca(OH2)n (n ) 1) in the sodalite cages.
1H/27Al TRAPDOR NMR demonstrated that the water mol-
1
ecules are in close proximity to aluminum atoms. The H
resonances from CaOH+ groups in the sodalite cages resonate
at 2.1-2.8 ppm and are most easily seen in the 1H/27Al
TRAPDOR spectrum when 27Al irradiation was applied. A
concentration of approximately 16 Brønsted acid sites/unit cell
was determined by adsorbing TMP molecules on the sample
activated at 500 °C. Only a small number of Brønsted acid
sites were detected, when samples were carefully calcined in a
1
vacuum. No Lewis acidity was observed. A H resonance at
0.7 ppm observed in the spectrum of CaY calcined at 200 °C is
(20) Massiot, D.; Bessada, C.; Coutures, J. P.; Taulelle, F. J. Magn.
Reson. 1990, 90, 231.
assigned to CaOH+ groups in the supercages.
(21) (a) Robbins, R.; Ramamurthy, V. J. Chem. Soc., Chem. Commun.
1997, 1071. (b) Li, X.; Ramamurthy, V. J. Am. Chem. Soc. 1996, 118,
10666. (c) Robbins, R.; Crompton, E.; Ramamchandran, S.; Ramamurthy,
V. Unpublished results.
(22) (a) Pitchumani, K.; Ramamurthy, V. J. Chem. Soc., Chem. Commun.
1996, 2763. (b) Jayathirtha Rao, V.; Prevost, N.; Ramamurthy, V.; Kojima,
M.; Johnston L. J. J. Chem. Soc., Chem. Commun. 1997, 2209.
(23) For carbocation formation in solution: (a) Gold, V.; Hawes, B.
W. V.; Tye, F. L. J. Chem. Soc. 1952, 2167. (b) Gold, V.; Tye, F. L. J.
Chem. Soc. 1952, 2172. (c) Evans, A. G.; Jones, N.; Jones, P. M. S.; Thomas,
J. H. J. Chem. Soc. 1956, 2757. (d) Evans, A. G.; Jones, P. M. S.; Thomas,
J. H. J. Chem. Soc. 1957, 104. (e) Grace, J. A.; Symons, M. C. R. J. Chem.
Soc. 1959, 958.
(24) For carbocation formation on a silica-alumina surface, see the
following: (a) Leftin, H. P.; Hall W. K. J. Phys. Chem. 1960, 64, 382. (b)
Leftin, H. P. J. Phys. Chem. 1960, 64, 1714. (c) Leftin, H. P.; Hall, W. K.
J. Phys. Chem. 1962, 66, 1457. (d) Leftin, H. P.; Hobson, M. C.; Leigh, J.
S. J. Phys. Chem. 1962, 66, 1214. (e) Rooney, J. J.; Pink, R. C. Trans.
Farady Soc. 1962, 1632. (f) Hall, W. K. J. Catal. 1962, 1, 53.
(25) (a) Rooney, J. J.; Hathaway, B. J. J. Catal. 1964, 3, 447. (b) Dollish,
F. R.; Hall, W. K. J. Phys. Chem. 1965, 69, 4402.
The number of Brønsted acid sites present within CaY is
controlled by the mode of activation. This determines the
activity of CaY and, in turn, the nature of the products obtained.
For the reactions studies in this paper, the Brønsted acid sites
reacted stoichiometrically when present in high concentrations,
resulting in the reduction of the olefin. At lower concentrations,
the Brønsted acid sites acted catalytically to affect isomerization
reactions. Finally, this paper shows that a combined zeolite
characterization and product reactivity study of the same
materials, activated under identical conditions, can be very
helpful in rationalizing the product distributions, especially in
reactions where both a wide range of products are possible and
with zeolites whose properties can vary significantly with
activation conditions.
Acknowledgment. The National Science Foundation Na-
tional Young Investigator program (Grant DMR-9458017) is
thanked for their partial support of this research. The solid-
state NMR spectrometer was purchased with a grant from the
National Science Foundation (CHE-9405436). V.R. thanks the
Division of Chemical Sciences, Office of Basic Energy Sciences,
Office of Energy Research, U.S. Department of Energy, for
support of this program.
(26) Hirschler, A. J. Catal. 1963, 2, 428.
(27) Fornes, V.; Garcia, H.; Jovanvic, S.; Marti, V. Tetrahedron 1997,
53, 4715.
(28) Reviews: (a) Leftin, H. P.; Hobson, M. C. AdV. Catal. 1963, 14,
115. (b) Terenin, A. AdV. Catal. 1964, 15, 227.
(29) (a) Al-Ekabi, H.; Kawata, H.; de Mayo, P. J. Org. Chem. 1988,
53, 1471. (b) Shida; T.; Hamill, W. H. J. Chem. Phys. 1966, 44, 4372. (c)
Brede, O.; Bos, J.; Helmstreet, W.; Mehnert, R. Radiat. Phys. Chem. 1982,
19, 1. (d) Mizuno, K.; Tamai, T.; Hashida, I.; Otsuji, Y.; Kuriyama, Y.;
Tokumaru, K. J. Org. Chem. 1994, 59, 7329. (e) Konuma, S.; Aihara, S.;
Kuriyama, Y.; Misawa, H.; Akaba, R.; Sakuragi, H.; Tokumaru, K. Chem.
Lett. 1991, 1897.
(30) (a) Bartlett, P. D., Ed. Nonclassical Ions; W. A. Benjamin: New
York, 1965. (b) Pocker, Y. In Molecular Rearrangements; de Mayo, P.,
Ed.; Interscience: New York, 1963; Part 1, pp 1-25.
(31) At this stage, both 17 and 19 could account for the long λ
absorption. Mechanism provided in Scheme 6 accounts for the formation
of both 17 and 19 as well as for the isolated products 2, 3, 20, and 24.
References and Notes
(1) (a) Pitchumani, K.; Joy, A.; Prevost, N.; Ramamurthy, V. J. Chem.
Soc., Chem. Commun. 1997, 127. (b) Pitchumani, K.; Lakshminarasimhan,
P. H.; Prevost, N.; Corbin, D. R.; Ramamurthy, V. J. Chem. Soc., Chem.
Commun. 1997, 181. (c) Pitchumani, K.; Lakshminarasimhan, P. H.; Turner,
G.; Bakker, M. G.; Ramamurthy, V. Tetrahedron Lett. 1997, 371. (d)