Effects of the properties of SO4/ZrO2 solid catalysts on the products of transformation and reaction mechanism of R-(+)-limonene diepoxides

Article abstract of DOI:10.1016/j.molcata.2007.01.005

Transformations of limonene diepoxides on solid SO₄/ZrO₂ catalysts in dried methylene chloride at room temperature were investigated by varying the amounts of supported sulfate ions and zirconia polymorphs. A method for estimating acid centers from DRIFTS spectra of adsorbed pyridine on the solid catalyst in a dried methylene chloride solution has been developed. It is shown that the mechanism of diepoxide transformations is changed by varying the type and amount of acid centers. At a 0.9-3% level of sulfation, Lewis centers are mainly responsible for the initial cleavage of the 8,9-epoxy group of limonene diepoxides (route B). At higher contents of sulfate ions (up to 30 wt.% SO₄), the key transformations proceed with cleavage of the 1,2-epoxy group (route A) and involve the Br?nsted centers. The type of support affects the structural features of the sulfate ions, which govern the rearrangement of limonene diepoxides by route A or B. Transformation of limonene diepoxides on the Lewis centers formed on sulfated alumina follows mainly route A. The type of acid centers also affects the stereochemical composition of the product mixture.

Full text of DOI:10.1016/j.molcata.2007.01.005

Journal of Molecular Catalysis A: Chemical 269 (2007) 72–80
Effects of the properties of SO4/ZrO2 solid catalysts
on the products of transformation and reaction
mechanism of R-(+)-limonene diepoxides
O.V. Salomatinaa,∗, T.G. Kuznetsovab, D.V. Korchaginaa, E.A. Paukshtisb,
E.M. Morozb, K.P. Volchoa, V.A. Barkhasha, N.F. Salakhutdinova
a N.N. Vorozhtsov Novosibirsk Institute of Organic Chemistry, Siberian Branch of the Russian Academy of Sciences,
Acad. Lavrentiev Ave. 9, Novosibirsk 630090, Russia
b G.K. Boreskov Institute of Catalysis, Acad. Lavrentiev Ave. 5, Novosibirsk 630090, Russia
Received 4 July 2006; received in revised form 27 December 2006; accepted 4 January 2007
Available online 11 January 2007
Abstract
Transformations of limonene diepoxides on solid SO4/ZrO2 catalysts in dried methylene chloride at room temperature were investigated by
varying the amounts of supported sulfate ions and zirconia polymorphs. A method for estimating acid centers from DRIFTS spectra of adsorbed
pyridine on the solid catalyst in a dried methylene chloride solution has been developed. It is shown that the mechanism of diepoxide transformations
is changed by varying the type and amount of acid centers. At a 0.9–3% level of sulfation, Lewis centers are mainly responsible for the initial
cleavage of the 8,9-epoxy group of limonene diepoxides (route B). At higher contents of sulfate ions (up to 30 wt.% SO4), the key transformations
proceed with cleavage of the 1,2-epoxy group (route A) and involve the Brønsted centers. The type of support affects the structural features of the
sulfate ions, which govern the rearrangement of limonene diepoxides by route A or B. Transformation of limonene diepoxides on the Lewis centers
formed on sulfated alumina follows mainly route A. The type of acid centers also affects the stereochemical composition of the product mixture.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Limonene diepoxides; Sulfated zirconium oxide; Rearrangements; Mechanism; Acid catalysis; Alumina
1. Introduction
to activation of known transformations, but also occasionally
changes the reaction route compared to routes in homogeneous
media [3].
The epoxy ring readily formed by oxidation of double bonds
exhibits high reactivity in acidic and basic media [1]; this is
part of many biologically active compounds, both natural and
synthetic, and an important building block in organic synthesis
[2]. Because of their unique structure and biological activity,
epoxy terpenes are very important substrates for chemical syn-
thesis; these are inexpensive and readily available compounds.
Reactions with these compounds used as substrates afford prod-
ucts from various classes of organic compounds. Using zeolites,
clays, and other solid acids as catalysts of acid transformations
of terpenes and their epoxy derivatives not only improves the
ecological characteristics of processes and lowers the barriers
As is known, dissolution of limonene diepoxides 1a,b
in homogeneous acid media leads to reactions necessarily
involving an external nucleophile. The cation center initially
formed during cleavage of the protonated epoxy ring either
reacts immediately with the external nucleophile, giving the
products of independent cleavage of epoxy rings 2a,b–3a,b
[4,5] or undergoes intramolecular cyclization, forming the
6-oxabicyclo[3.2.1]octane cation trapped by the external nucle-
ophile to give compounds 4a,b–5a,b [5] (Scheme 1). Thus
dissolving diepoxides 1a,b in 1N HCl in ether leads to cleav-
age of the 1,2-epoxy ring, forming hydrochloride 2a,b. In the
acetone–water–sulfuric acid system, the 1,2-epoxy ring is con-
verted into diol, and the 8,9-epoxy group reacts with the acetone
molecule to form 1,3-dioxolane ring 3a,b. Transformations of
limonene diepoxides 1a,b in formic and trifluoroacetic acids
∗
Corresponding author. Tel.: +7 3833308870; fax: +7 3833309752.
E-mail address: ana@nioch.nsc.ru (O.V. Salomatina).
1381-1169/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2007.01.005

O.V. Salomatina et al. / Journal of Molecular Catalysis A: Chemical 269 (2007) 72–80
73
Scheme 1.
lead to 6-oxabicyclo[3.2.1]octane derivatives 4a,b–5a,b. Iso-
merization on solid acid catalysts (zeolite-␤, askanite-bentonite
clay, clay K-10, and SO4/TiO2) does not involve an external
nucleophile and leads to a wide spectrum of optically active bis-
and tricyclic oxygen-containing compounds (Scheme 1) [6]. The
qualitative and quantitative compositions of the reaction mixture
depend on the type of the solid catalyst used. On clays, com-
pound 7 was the major product, while 6, 8a,b, and 9 formed in
smaller amounts (ratio of 6, 7, 8a,b, and 9 was ∼4:7:2:3, respec-
tively, GLC data). On zeolite and sulfated titania, the reaction
gave 6, 7, and 10 (product ratio was ∼2:2:5 and ∼3:1:2, respec-
tively, GLC), the major product being either compound 10 or
compound 6 (GLC).
tion of the oxygen atom(s) to the Zr centers (reducing the S O
bond order) [24–25]. Yamaguchi [26] assumed that the sulfate
chelates one Zr atom with a double oxygen bond, attracting elec-
trons and making it a strong Lewis acid. Clearfield et al. [27]
suggested that sulfur bridges two zirconium atoms and that by
water sorption Lewis acid sites are converted to Brønsted acid
sites. In an HRTEM study, Benaissa et al. [28] observed two- or
threefold coordination of sulfate groups to zirconia crystallites.
Interpretation of the acidic properties of sulfated zirconia is fur-
ther complicated by the dependence of the nature of this material
on the preparation methods and conditions [29–32]. The sulfate
ions, adding during the synthetic procedure, stabilize tetragonal
zirconia, but increase the particle size of the phase [14].
It is important that contrary to homogeneous media, solid
acid catalysts promote biomimetic mechanisms of the transfor-
mations of limonene diepoxides. For example, isobottrospicatol
9 was previously obtained by the action of Streptomyces bot-
tropensis on carveol [7], compound 10 was isolated from the
aerial parts of Haplopappus multifolius [8], and tricyclic diethers
8a,b possess the framework of paeonimetaboline [9].
The significant effect of the nature of the solid acid catalyst
on the composition of the product mixture and product ratio
in transformations of limonene diepoxides 1a,b prompted us to
investigate the catalyst structure effects on the reaction route to
gain more profound understanding of the processes and to study
possible effects of some characteristics of the solid catalyst on
the composition of the reaction mixture.
Sulfated zirconia showed remarkable activity in reactions
not catalyzed by conventional solid catalysts, e.g., in skele-
tal isomerizations of alkanes, alkane alkylations with alkenes,
low-temperature crackings of polyolefins, Friedel–Crafts type
alkylations, citronellal cyclizations, and benzoic acid etherifi-
cations [10–19]. The high catalytic activity of sulfated zirconia
was attributed to superacidity of this compound. However, some
authors claimed that sulfated zirconia has no superacidity or
that catalytic activity is not related to superacidity [20–23]. The
state of the surface of zirconia polymorphs may influence the
nature of the adsorbed sulfate, leading to ionic or covalent sul-
fates [24–28]. The covalent character of the sulfate depends on
the degree of surface dehydration and on the type of coordina-
The role of different types of acidic center and polymorphous
modifications of zirconia in the reactions is also the source of
controversy. Thehigheractivityofpseudocubicsulfatedzirconia
in n-butane isomerization compared with tetragonal and espe-
cially monoclinic modifications cannot be attributed to changes
in surface functionalities, but is believed to be the consequence
of better stabilization of the high-symmetry crystalline form
(due to the more favorable distribution of sulfates in regular and
defective parts of crystals) [12]. Both Lewis and Brønsted active
acid centers were claimed to be important for the reaction. For
nanosizedzirconiaappliedtoalumina-silicatelayersofmontmo-
rillonite(Zrpillaredclay)andconsistingpredominantlyofsingle
Zr4 tetramers, sulfation at low loading starts with formation of
isolated sulfate bound with Zr4+ cations [13]. The increased
acidity (Brønsted centers) is associated with polymerized sulfate
species formed at high loading and guaranteeing higher activ-
ity in n-hexane isomerization and isopropanol dehydration. The
Brønsted centers generated by loading excess sulfate are active
in skeletal isomerization, while the Lewis centers of submono-
layer sulfate, covering the surface, catalyze Friedel–Crafts type
alkylation of benzene with benzyl chloride in the liquid phase
[11]. The presence of both Lewis and Brønsted sites is essential
to cyclization of citronellal to isopulegol [18].
This controversy in determining active reaction centers may
be associated not only with the complexity of reaction mecha-
nisms or the effects of the synthetic procedure of catalysts, but
also with methods for evaluating the acid centers of the cata-

74
O.V. Salomatina et al. / Journal of Molecular Catalysis A: Chemical 269 (2007) 72–80
Table 1
Chemical composition, specific surface area (SSA), and number of acidic centers for sulfated catalysts (wt.% SO4)
Sample number
Chemical composition
SSA (m2/g)
Number of acidic centers (a.u.)
Lewis type
Brønsted type
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
Sample 6
Sample 7
Sample 8
Sample 9
Sample 10
0.9% SO4/90% M + 10% T-ZrO2
0.9% SO4/C-ZrO2
63
60
n.d.a
n.d.
1.6
n.d.
0.7
n.d.
0
n.d.
3.7
0
n.d.
n.d.
0
n.d.
2.8
n.d.
4.5
n.d.
0
0.9% SO4/60% T + 40% M-ZrO2
1.8% SO4/60% T + 40% M-ZrO2
3.0% SO4/60% T + 40% M-ZrO2
9.0% SO4/60% T + 40% M-ZrO2
15.2% SO4/60% T + 40% M-ZrO2
30.3% SO4/60% T + 40% M-ZrO2
3.0% SO4/␥-Al2O3
145
140
130
110
100
81
189
240
0.9% SO4/Zr PL
2.0
a
n.d.: not determined.
lysts. According to quantum chemical data, the strength of acid
centers in sulfated zirconium dioxide is close to that in sulfu-
ric acid, but weaker than that in zeolites [33,34]. The data of the
tracer method confirm that the strength of the acid centers of sul-
fated zirconium dioxide approaches the strength of these centers
in 100% sulfuric acid [35]. However, varying the color of the
Hammett indicator may affect the chemical nature and the envi-
ronment of the solid surface, resulting in detection of properties
that are far from the working state of the catalyst. Moreover,
it is generally difficult to derive exact quantitative data from
the color of the indicator. To determine the acidic properties
of sulfated zirconium dioxide one can also employ gas-phase
adsorption of ammonia and pyridine [11–13,18]. However, ther-
mally programmed desorption of ammonia resulted in peaks
observed at high temperatures well above the reaction temper-
atures. Ammonia adsorbed on a pure support may be removed
by preliminary treatment of the surface with steam, significantly
reducing the temperature of the desorption peak [11]. In the case
of pyridine, the surface centers may be detected from the charac-
teristic IR bands of various adsorbed complexes. In the region of
adsorbed complexes, however, IR spectra depend significantly
on the desorption temperature. Preliminary sample training in
vacuum before adsorption is a substantial disadvantage com-
mon to all of these methods because it can drastically change
the composition of the surface.
In this work we have studied for the first time the transfor-
mations of limonene diepoxides in the presence of methylene
chloride on ZrO2 solid catalysts at room temperature while
varying the structure of zirconia and activating the surface with
supported sulfate ions. The following modifications of highly
disperse bulk zirconia are known: monoclinic (M), tetragonal
(T), and partially stabilized (e.g., by calcium cations) low-
temperature cubic (C) forms [36,37]. We used three zirconia
samples: (90% M + 10% T), (60% T + 40% M), and (C + 5%
CaO) with particle size varying from 5 to 10 nm. The specific
surface of the starting pure support was 74, 150, and 72 m2/g for
the prevalent M, T, and C modifications of zirconia, respectively.
Nanosized zirconia (particle size 1.0–1.5 nm), stabilized in the
interlayer space (gallery height ∼ 0.7 nm) of layered montmoril-
loniteclay(ZrPL)(whichdiffersfrombulkZrO2 initsproperties
[38]), was used to estimate the size effect of zirconia particles.
The content of the sulfate ions in the catalysts varied from 0.9 to
30 wt.% SO4. The acidic (Lewis and Brønsted) properties of sul-
fated zirconias were investigated by analyzing the characteristic
bands of pyridine in the DRIFT spectra after pyridine adsorption
from a methylene chloride solution at room temperature. This
method was elaborated in order to be able to compare the acidic
properties and activity of the catalysts under similar conditions.
2. Results
2.1. Structural effects of sulfated zirconia
On pure zirconium-containing supports, no transformations
of limonene diepoxides 1a,b have been observed. For sulfated
samples, the activity and selectivity with respect to reaction
products depend on the modification of zirconia and on the
amount of supported sulfate ions. Comparative experiments on
the structure effects of sulfated zirconias on the transformations
of limonene diepoxides 1a,b were carried out on samples 1–3
containing 0.9 wt.% SO4 (Table 1). For all these samples, the
reaction formed the same products: cyclic alcohol 9 as the major
product (Scheme 2 and Table 2) together with minor amounts of
tricyclic diethers 6 and 7 (ratio of 9:6:7 is ∼10:1:1, GLC). How-
ever, the activity depends on the zirconia polymorphs. Thus after
24 h, conversion of the starting diepoxide was 60% for sample 1,
40% for sample 2, and 100% for sample 3. The highest activity
of sample 3 may be attributed to the specific features of sulfate
ions on zirconia supports: catalysts based on pseudocubic or
monoclinic zirconia are less active than catalysts based on the
tetragonal phase. The specific surface area (SSA = 150 m2/g),
as well as the degree of disordering and surface hydroxylation,
increased with the proportion of the tetragonal phase (IR and
Raman spectral data) [39]. This type of support containing a
mixture of the T and M modifications of zirconia was chosen
for experiments with varied contents of sulfate ions.
2.2. Effects of the content of sulfate ions
Catalysts containing 0.9–30.3 wt.% SO4 were prepared.
Table 1 gives a list of catalyst products. For catalysts prepared
by wetness impregnation with an (NH4)2SO4 solution and espe-

O.V. Salomatina et al. / Journal of Molecular Catalysis A: Chemical 269 (2007) 72–80
75
Scheme 2.
cially for catalysts with over 3 wt.% SO4, the specific surface
decreased. When sulfate ions are present in high contents, their
location along the domain boundaries probably favors the for-
mation of large particles of zirconia, leading to two-dimensional
sulfates. Zirconia particles were also observed to enlarge after
significant amounts of sulfate ions were added at the stage of
synthesis [14].
Table 2 lists data on the transformations of diepoxides 1a,b
on sulfated zirconia with different concentrations of sulfate ions
(from 0.9 to 30.3 wt.% SO4). Increased contents of sulfate ions
(more than 0.9 wt.% SO4, samples 4–8) led to higher catalyst
activity in the transformations of diepoxides 1a,b; 100% con-
version of these compounds was observed even after 5 h (versus
24 h for sample 3) of reaction at room temperature. Isomeriza-
tion of diepoxides 1a,b on sample 3 containing 0.9 wt.% SO4
selectively formed isobottrospicatol 9 and minor amounts of
compounds 6 and 7 (Scheme 2). At higher sulfate concentrations
(1.8–3.0 wt.% SO4), the composition of the reaction mixture
gradually became more complex: the relative content of com-
pound 9 decreased, while the contents of compounds 6 and 7
increased, and new products appeared in the reaction mixture:
8a,b, 10, 11, 12, and 13. For catalysts containing more than
9% sulfate ions, compound 9 was present in minor quantities
if at all; the ratio of other compounds did not change signifi-
cantly. The total peak areas of the products are independent of
the composition of the zirconia-based catalysts (Table 2), indi-
cating that the catalysts generally have the same level of activity
irrespective of the sulfate ion content. The contribution of other
Table 2
Product ratios in transformations of limonene diepoxides 1a,b on sulfated catalysts
ꢀ
Catalysta
Product ratiob
Selectivityd (%)
c
6
7
8a,b
9
10
11
12, 13
Route A
Route B
Sample 3e
Sample 4
Sample 5
Sample 6
Sample 7
Sample 8
Sample 9
Sample 10
1.1
1.6
1.7
1.7
1.8
1.9
0.4
3.9
1.1
2.5
2.5
2.5
2.5
2.4
2.9
3.1
–
10.4
8.7
4.3
0.4
–
–
–
3.6
–
–
1.7
3.4
3.7
3.6
–
–
–
12.6
13.7
14.0
12.4
13.0
13.1
18.8
15.7
9
13
38
61
65
66
94
51
91
87
62
39
35
34
6
0.7
1.9
2.0
2.1
2.1
–
0.2
1.3
1.3
1.8
1.8
–
0.6
1.1
1.1
1.3
–
15.5
2.8
1.0
0.9
0.4
49
a
Reactions conducted in methylene chloride for 5 h, ratio of starting diepoxides 1a:1b = 3:2, ratio of solid catalyst:diepoxides = 5:2.2 (by mass).
Peak areas given according to GLC data in reaction mixtures with an internal standard whose peak area is taken for unity.
The sum of all peak areas on the chromatogram.
Route A, compounds 6, 10–13; route B, compounds 7–9.
Reaction conducted for 24 h.
b
c
d
e

76
O.V. Salomatina et al. / Journal of Molecular Catalysis A: Chemical 269 (2007) 72–80
types of reactions (e.g., polymerization, oxidation, etc.) did not
change with the chemical composition of the catalyst and was up
to 10% of the total amount of diepoxides 1a,b used (evaluated
from the weight of the compounds separated from the catalyst).
Major differences between the catalysts are reflected by product
selectivity and may be caused by varying the structure of active
sulfate ions.
Scheme 2 shows the possible mechanism of transformations
of diepoxides 1a,b [6]. The reaction possibly starts with cleav-
age of the protonated 1,2-epoxy group of diepoxides 1a,b (route
A, compounds 6, 10–13 as products) or of the protonated 8,9-
epoxy group (route B, compounds 7–9 as products). Analysis of
thedependenceoftheproductratioonthecatalystinthetransfor-
mations of limonene diepoxides 1a,b (Table 2) revealed that the
contribution of reaction route A or B depends on the concentra-
tion of the sulfate ions. For catalysts with low contents of sulfate
ions, the major reaction products were the compounds that corre-
sponded to the initial cleavage of the 8,9-epoxy group (route B),
while increased sulfate ion concentration led to increased con-
tents of products formed by the initial cleavage of the 1,2-epoxy
group (route A). Thus activation of polyfunctional terpenoid
1a,b is rather sensitive to the structure of active centers on the
surface.
The relative distribution of isomerization products within
groups of compounds formed by different routes also depends
on the concentration of the sulfate ions. While compound 6 was
present in the reaction mixture in pronounced (but moderate,
8–15%) quantities at any concentration of the sulfate ions, 2-
oxabicyclo[3.3.1]nonanes 10–13 formed in significant amounts
only on the sample containing 9.0% supported sulfate ions.
At higher sulfate concentrations, the relative contents of these
compounds in the mixture were practically independent of the
catalyst composition.
Compound 9 with a 2-oxabicyclo[3.2.1]octane type frame-
work formed as the major product on samples 3–5, but vanished
from the reaction mixture completely when the samples con-
tained more than 9.0% supported sulfate ions. At the same time,
tricyclic diether 7 obtained by intramolecular heterocyclization
in cation IV (Scheme 2) structurally related to cation III that
forms compound 9 (the only difference being the position –
endo or exo – of the hydroxymethylene group) was present on
all sulfated catalysts of the series. Thus varying the concentra-
tion of supported sulfate ions is critical not only to the site at
which the cation center appears, but also to the stereochemical
structure of products.
tains 20 wt.% ZrO2. When calculated based on bulk zirconia,
the content of sulfate ions in this sample is 4.5 %, which is inter-
mediate between the contents of these ions in samples 5 and 6.
At the same time, percent of compound 6 in sample 10 is larger,
which may be due to the change in the structure of sulfate ions
on the surface of nanosized zirconia in pillared clay. Since the
latter preferably contains discrete Zr4 tetramer units, formation
of two- and three-dimensional sulfates on deposited sulfate ions
was not likely as was the case with bulk ZrO2 [40]. While pure
pillared clay containing nanosized zirconia shows no activity
in the transformation of limonene diepoxides 1a,b, sulfated pil-
lared clay is active in the reaction and gives product distribution
other than that in montmorillonite clays. On sample 10, com-
pounds 6, 7 and 9 formed in approximately equal proportion,
while on montmorillonite clay compound 7 was the dominant
product [6]. Moreover, transformation of limonene diepoxides
1a,b on sulfated pillared clay proceeds with formation of new
compounds 10–13. Formation of these compounds may be the
result of specific activation of limonene diepoxides 1a,b on this
type of sulfated nanomaterial.
Using alumina instead of zirconia as support for sulfate
ions (sample 9) led to substantial changes in the product ratio
(Table 2). Compounds formed by route A were the major prod-
ucts. At the same time, on pure alumina no transformations of
limonene diepoxides 1a,b have been detected. The product ratio
onsample9wassimilartothatobservedpreviouslyonthezeolite
catalyst [6]. Consequently, the properties of acid centers on the
surface of sulfated alumina are similar to the properties of these
centers in zeolites, while sulfated zirconia creates other types of
acid center. Compounds 10–13 were not detected in the reaction
mixture on sample 9 either. These differences may be caused by
differences in the structure and properties of sulfates supported
to the surface of ␥-Al2O3 and ZrO2. While aluminum cations
preferably have octahedral surroundings of oxygen atoms, zir-
conium cations each lies at the center of a cube with coordination
number 7 or 8 depending on the structural modification.
Thus the reaction route of limonene diepoxides 1a,b on the
catalysts under study is largely determined by the chemical
nature of sulfated active centers.
2.3. Acidity properties
As is known, Lewis and/or Brønsted acidity can easily be
determined by using pyridine (Py) adsorption/desorption at RT
[41]; IR spectra show well-resolved specific bands that may be
attributedeithertoLewisacidsites(19bmodeof{Py-L}species,
≈1445 cm−1) or to Brønsted acid sites (19b mode of {Py-B}
species, ≈1540 cm−1). To find out if the differences in the cat-
alytic activity exhibited by different sulfated materials are due
to the acidic character of these compounds, we carried out Py
adsorption experiments under conditions of limonene diepoxide
transformations (details of the method are given below). The
difference DRIFT spectra of the samples obtained by subtract-
ing the spectra of {dried methylene chloride + sample} from the
spectra of {dried methylene chloride + Py + sample} are shown
in Fig. 1. One can clearly see the bands at 1440 and 1540 cm−1
ascribed to Lewis and Brønsted acidity centers; the intensity of
Reactions of limonene diepoxides 1a,b with sulfated Zr pil-
lared clay containing 0.9 wt.% SO4 (sample 10) proceeded more
readily than reactions with sample 3 having the same amount of
supported sulfate ions. Full conversion of limonene diepoxides
1a,b on sample 10 was revealed after 5 h of reaction versus 24 h
for sample 3. At the same time, transformations of limonene
diepoxides 1a,b over sulfated pillared clay containing smaller
zirconia particles (sample 10; 1.0–1.5 nm versus 7.5 nm for 60%
T + 40% M) gave a wide range of products with practically equal
distributions between routes A and B (Table 2). As a matter
of fact, the product ratio is intermediate between the ratios for
samples 5 and 6 prepared from bulk zirconia. Pillar clay con-

O.V. Salomatina et al. / Journal of Molecular Catalysis A: Chemical 269 (2007) 72–80
77
Fig. 1. Difference DRIFT spectra of adsorbed pyridine at RT on the initial
support of 60% T + 40% M-ZrO2 and sulfated catalysts (see note in Table 1).
these bands depends on the chemical composition of the cata-
lyst (Table 1). The intensities of other bands, attributed to the
overlap of H2O, PyH+, and Py-L absorption bands, are not dis-
cussed here. It may be assumed that the observed bands of the
{Py-B} and {Py-L} species belong to the rather strong acidic
forms on the surface of the catalyst. For the initial support con-
taining 60% T + 40% M-ZrO2, only {Py-L} complexes were
observed (Fig. 1), which appeared due to the presence of the
surface Zr4+ cations with an uncrowded coordination sphere. In
spite of the high degree of hydroxylation [39], {Py-B} com-
plexes with surface OH groups did not form on this type of
support because of their neutral or basic nature. With small addi-
tions of sulfate ions, new Lewis centers can appear due to the
interaction with Zr4+ cations having an uncrowded coordina-
tion sphere. Isolated sulfate ions may change the space near the
Zr4+ cations, producing complex active centers involving sur-
face hydroxyls. When the addition of sulfate ions increases, the
amounts of the {Py-L} species reach a maximum at 0.9 wt.%
SO4, while the {Py-B} complexes become visible in the DRIFT
spectra at 3 wt.% SO4, increasing at higher sulfate concentra-
tions. The dependence of the number of acidic centers referred
to the SSA on the amount of the supported sulfate ions is shown
in Fig. 2a. The observed dependences could be explained by
the following processes. When sulfate ion additions are small
(0.9% SO4), the ions can form new types of Lewis center with
Zr4+ cations having an uncrowded coordination sphere. The
content of this type of defect on the surface of zirconia does
not exceed a few percent of the monolayer cover. In this case,
Lewis centers can appear due to adsorption on the Zr4+ cations
having an uncrowded coordination sphere or in the vicinity of
these cations, producing complex active centers. At this level of
supported sulfate ions, they occupy ≈1.5% of the surface area
Fig. 2. Numbers of surface Lewis (1) and Brønsted (2) acidic centers per m2 (a)
and selectivity of transformations of limonene diepoxides 1a,b, route A (1) and
route B (2) (b) vs. the amounts of sulfate groups on the surface of ZrO2.
are detected, the SSA diminishing noticeably. Thus an “island”
model of sulfate centers may be proposed based on these data.
After addition of sulfate ions, Brønsted centers appear around
Lewis type centers, probably forming along the microdomain
boundary and leading to sintering.
For Zr PL containing nanosized particles of ZrO2 smaller
than those of bulk zirconias studied here, sulfation leads to only
Brønsted centers detected by Py adsorption (Fig. 1). As a matter
of fact, pure Zr PL is characterized by high density of Lewis
type Zr4+ cations having an uncrowded coordination sphere and
detected by CO adsorption; it also has high density of surface
hydroxyls originating from Zr4 tetramers [38], while the acidic
sites of initial montmorillonite are suppressed by intercalation of
the tetramers. According to analysis of low-temperature adsorp-
tion of nitrogen [42], the interlayer space in pillared clay is filled
with isolated Zr4 tetramers and sheetlike Zr8 species separated
by ∼0.7–1.0 nm and bound with montmorillonite layers. After
sulfation, isolated sulfate species that are Lewis and/or Brønsted
centers by nature could appear. Due to the high density of the
surface hydroxyls, Lewis centers are accessible for CO (but not
Py) adsorption.
of the support, the occupation area SSO ≈ 0.31 nm2 included
4
[11]; the SSA of sample 3 practically remains unchanged com-
pared to the area of the initial support. The amount of 3 % SO4
might increase the number of Lewis centers as the sulfate ions
occupy only ≈3% of the total area. In reality, however, Lewis
centers decrease in number, and Brønsted centers appear. The
SSA of sample 5 also decreases relative to that of the initial sup-
port. At higher contents of sulfate ions, only Brønsted centers
After adsorption of Py by sulfated alumina, only Lewis cen-
ters are detected (Fig. 1). It is well known that Py does not form
Brønsted complexes with hydroxyls of alumina. After 3 wt.%

78
O.V. Salomatina et al. / Journal of Molecular Catalysis A: Chemical 269 (2007) 72–80
SO4 had been applied to alumina, the SSA of the catalyst did
not change as was the case with sulfated zirconia (Table 1, sam-
ple 9), which could be evidence of the formation of isolated
sulfated species. In fact, it is also well known that the number
of centers with an uncrowded coordination sphere on alumina
is higher than that on zirconia.
polarimeter. The element composition of the products was deter-
mined from high-resolution mass spectra recorded on a Finnigan
1
MAT 8200 instrument. H and 13C NMR spectra were mea-
sured on an AM-400 Bruker spectrometer (operating frequency
1
400.13 MHz for H and 100.61 MHz for 13C) for CDCl3 or
(1:1) CCl4–CDCl3 solutions of the substances. The signal of
chloroform was used as an internal standard (δH 7.24 ppm, δC
76.90 ppm). The structure of the compounds was determined
3. Discussion
1
by NMR from proton spin–spin coupling constants in H–1H
Thus the transformations of limonene diepoxides were gradu-
ally changed by varying the properties of sulfated zirconia-based
catalysts. It has been shown for the first time that the mechanism
of diepoxide transformations may be changed by varying the
concentration and type of acid centers in nanosized SO4/ZrO2
(Fig. 2). At a low level of 0.9–3% of sulfation, Lewis centers
preferably formed, which are responsible for the initial cleav-
age of the 8,9-epoxy group of limonene diepoxides (route B).
With higher amounts of supported sulfate groups, Brønsted cen-
ters are the major species on the surface, giving a variety of
products of initial cleavage of both 8,9- (route B) and 1,2-epoxy
groups (route A). The type of support affects the structural fea-
tures of the sulfate ions responsible for activation of limonene
diepoxides by route A or B. Activation of limonene diepoxides
on Lewis centers formed on sulfated alumina follows route A.
The concentration and type of acid centers also affect the
stereochemical composition of the product mixture. In the pres-
ence of Lewis centers on zirconia, the compound that formed
via cation III with an exo position of the hydroxymethyl group
was the major product. After Brønsted centers appeared along
with Lewis centers (for amounts of up to 5 wt.% SO4), this com-
pound vanished from the reaction mixture; however, the content
of the compound formed via intermediate cation IV with an
endo position of the hydromethyl group remained almost con-
stant. Evidently, the environment of the acid center along with
its type favors a certain conformation of the reactant molecule.
It has been shown that catalyst activity in transformations
of limonene diepoxides depends on the structural modification
of sulfated zirconia (T, M or C). For sulfated catalysts, activ-
ity changes in the series T > M > C. Optimum disordering and
hydroxylation of the T phase surface provides the highest activ-
ity of the sulfated catalysts. The lower activity of the sulfated
monoclinic versus tetragonal zirconia is thought to be caused
by the low symmetry of the strained monoclinic modification as
revealed by IR and DR-UV spectra [39], but not by the presence
of the trigonal oxide ions in the structure of monoclinic zirconia
[43].
double resonance spectra, and by analyzing 13C NMR spec-
tra using proton selective and off-resonance saturation spectra,
2D 13C–1H correlated spectroscopy on direct C–H constants
(COSY, 1JC,H 135 Hz), and 1D 13C–1H long-range J modulation
difference (LRJMD, JC,H 10 Hz).
4.2. Sample preparation
Zirconias with a prevalent monoclinic or tetragonal modifi-
cation were prepared from a 0.5 M solution of ZrOCl2 × 8H2O
by precipitation with aqueous ammonia at pH 10 with fur-
ther aging of the deposit. To prepare the dominant monoclinic
(M) modification, the deposit was aged by ultraviolet radiation
(high-pressure mercury lamp) for 12 h at room temperature.
After the deposit was washed with distilled water and then
dried and calcined at 500 ◦C for 4 h, it contained 90% M and
10% T (SSA = 74 m2/g) (XRD data). After the deposit was
aged at 100 ◦C for 53 h, washed with water, dried, and cal-
cined at 500 ◦C for 4 h, the sample contained 60% T and
40% M (SSA = 150 m2/g) (XRD data). The partially stabi-
lized (5 mol% CaO and 95 mol% ZrO2) low-temperature cubic
modification was prepared by the procedure of [18] using
polymerized organometal precursors from calcium nitrate and
ZrOCl2 × 8H2O; the final calcination temperature of the sample
was 500 ◦C after 6 h (SSA = 72 m2/g).
For
pillared
clay
containing
20 mass%
ZrO2
(SSA = 250 m2/g), details of synthesis are given in ref.
[19]. Natural montmorillonite clay was employed for synthesis.
The particle size of zirconia (1.0–1.5 nm) was estimated [19] by
assuming that two or three Zr4 tetramers entered the interlayer
space of pillared clay.
Zirconium-containing
samples
and
reactive
␥-
Al2O3 (SSA = 185 m2/g) were sulfated by incipient wetness
impregnation from a water solution of ammonium sulfate. The
final calcination temperature of the samples was 500 ◦C after
1 h.
ThespecificsurfaceareaofBET(SSA, m2/g)wasdetermined
from Ar thermal desorption data.
4. Experiment
4.3. Sample characterization
4.1. General
X-ray phase analysis of the samples was carried out using an
HZG-4C diffractometer (Cu K␣ radiation, flat monochromator)
in the 2θ range 1–70◦. Zirconia phases were identified by using
X-ray data files (PCDF): No. 37-1484 for the monoclinic phase,
No. 42-1164 for the tetragonal phase, and No. 27-0997 for the
low-temperature pseudocubic phase. The concentration of the
monoclinic phase mixed with the tetragonal modification was
Substrate purity and product composition were analyzed
by GLC using a Biochrome-1 chromatograph with a quartz
capillary column 13,000 mm × 0.22 mm, SE-54 phase. Flame
ionization detector, helium as carrier gas. The reaction prod-
ucts were separated by SiO2 column chromatography (CzFSR,
100–160 ␮m). Specific rotation was measured on a Polamat A

O.V. Salomatina et al. / Journal of Molecular Catalysis A: Chemical 269 (2007) 72–80
79
determined in the following way:
of chromatographic separations on a SiO2 column (hexane with
a gradient of diethyl ether from 0 to 80% as eluent) gave a mix-
ture of compounds 12 and 13 (0.025 g, yield 2%), 6 (0.05 g,
3%), 8a,b (0.03 g, 2%), 3 (0.07 g, 5%), 11 (0.01 g, 1%), and 10
(0.21 g, 14%). A sequence of separations on silica gave com-
pound 11 and a mixture of compounds 12 and 13. The low total
yield of the products is explained by the difficulties of separating
structurally related compounds.
−mon
mon
I
mon
+ I1 1 1
1 1 1
Cmon
=
I
1 1 1
−mon
tetr
1 1 1
+ I1 1 1 + I
The crystallite size was determined by the Scherrer equation
along the 1 1 1 line for the monoclinic modification, along the
1 0 1 line for the tetragonal form, and along the 2 2 0 line for
the cubic form: 10, 7.5, and 5.0 nm, respectively. Nanodisperse
zirconia in pillared clay was X-ray amorphous.
(1R,5R)-4,8-Dimethyl-2-oxabicyclo[3.3.1]non-3-en-8-ol
20
(11). [α] + 54.5 (c 1.1, CHCl3). MS, m/z: 168.12338
580
Diffuse reflectance infrared Fourier transform spectroscopy
(DRIFT) spectra of finely ground samples loaded in a small
cell (volume ≈ 0.2 ml) and placed in a chamber of DRS-8000
were recorded at room temperature in air at 2 cm−1 resolution
in the range 400–4000 cm−1 on a Shumadzu-8300 spectrometer.
Dried methylene chloride was added to the sample after stan-
dard treatment at 400 ◦C for 30 min; then the “wet” sample was
placed in a special cell of the spectrometer and the DRIFT spec-
trum was recorded. A solution of Py in dried methylene chloride
(0.12 mol/l), which ensures excess of Py over surface centers,
was added to the “fresh” portion of the sample, and a DRIFT
spectrumwasrecordedforthe“wet”sampleinthecellafter5and
10 min of adsorption. The difference spectra, obtained by sub-
tracting the spectra of dried methylene chloride + sample from
the spectra of dried methylene chloride + Py + sample, were used
for analyzing the acid centers of the catalysts. It has been shown
that the Py adsorption time does not influence the peak area of
the complexes. The total peak areas of the 1440 and 1540 cm−1
bandswereusedforestimatingtherelativeamountsofLewisand
Brønsted acid centers, respectively, in arbitrary units (Table 1).
1
[M+] C10H17O2. H NMR, δ: 1.24 (s, C11H3), 1.37 (d.d.d.d,
H7e, J7e,7a 13 Hz, J7e,6a 4 Hz, J7e,6e 2 Hz, J7e,1 1.5 Hz), 1.48
(d.d.d.d.d, H6e, J6e,6a 13 Hz, J6e,7a 5 Hz, J6e,7e 2 Hz, J6e,5 2 Hz,
J6e,9an 2 Hz), 1.51 (d, C10H3, J10,3 1.5 Hz), 1.58 (d.d.d, H7a, J
13 Hz, J7a,6a 13 Hz, J7a,6e 5 Hz), 1.59 (d.d.d.d, H9an, J9an,9syn
13 Hz, J9an,1 4 Hz, J9an,5 4 Hz, J9an,6e 2 Hz), 1.72 (d.d.d.d,
H6a, J 13 Hz, J6a,7a 13 Hz, J6a,7e 4 Hz, J6a,5 3.5 Hz), 2.03 (m,
H5), 2.14 (d.d.d, H9syn, J 13 Hz, J9syn,5 2.5 Hz, J9syn,1 1.5 Hz),
3.72 (d.d.d.d, H1, J1,9an 4 Hz, J1,5 1.5 Hz, J1,7e 1.5 Hz, J1,9syn
1.5 Hz) 6.15 (q, H3, J3,10 1.5 Hz). 13C NMR, δ: 75.76 (d, C1),
137.40 (d, C3), 109.53 (s, C4), 30.01 (d, C5), 24.93 (t, C6),
31.89 (t, C7), 72.53 (s, C8), 25.38 (t, C9), 16.94 (q, C10), 28.34
(q, C11).
Compounds 12 and 13 were not isolated as individual
compounds. NMR spectra were recorded for the ∼1:0.2
mixture of compounds 12 and 13: (1R,5R)-4,8-dimethyl-2-
oxabicyclo[3.3.1]non-3,7-dien(12). 1HNMR, δ:1.54(d, C10H3,
J10,3 1.5), 1.61 (d.d.d, H9syn, J9syn,9an 12.5, J9syn,1 2.5, J9syn,5 2),
1.75 (d.d.d, C11H3, J11,6 2.5, J11,6 1.5, J11,7 1.5), 1.85 (d.d.d.d,
ꢀ
H9an, J 12.5, J9an,5 4, J9an,1 3, J9an,6 1.5), 2.07 (d.d.d.d.d, H6,
J6,6 17, J6,7 5, J6,11 1.5, J6,5 1.5, J6,9an 1.5), 2.11 (m, H5), 2.18
ꢀ
ꢀ
(d.m, H6 , J 17), 4.26 (br.d.d, H1, J1,9ah 3, J1,9syn 2.5), 5.55
4.4. Transformations of diepoxides 1a,b
(d.d.q, H7, J7,6 5, J7,6 2, J7,11 1.5), 6.06 (q, H3, J3,10 1.5).
ꢀ
13C NMR, δ, м.д: 70.05 (d, C1), 135.22 (d, C3), 109.95 (s,
C4), 28.77 (d, C5), 31.26 (t, C6), 123.96 (d, C7), 131.76 (s,
C8), 28.06 (t, C9), 17.21 (q, C10), 21.44 (q, C11). (1R,5R)-4-
Limonene diepoxides obtained from R-(+)-limonene
(Aldrich, purity 97%, optical purity 98%) by the procedure of
[44] were employed; isomer ratio 1a:1b = 3:2 ([α] + 35.5 (c
20
580
1
3.5, CHCl3)) before the reaction the catalysts were calcinated
at 400 ◦C for 30 min. Dried methylene chloride (2 ml) was
added to the solid catalyst (0.05 g). Limonene diepoxides 1a,b
(0.022 g) were added with stirring to the resulting suspension.
The mixture was stirred for 5 h and then filtered through an
alumina layer (second degree of activity). Then a solution of
the internal standard (0.3 ml) was added to the reaction mixture.
The resulting reaction mixture was analyzed by GLC. A hexane
solution of dicyclopentadiene (10 mg of dicyclopentadiene
dissolved in 10 ml of hexane) was used as an internal standard
for chromatography. The analytical data are given in Table 2.
Compounds 11–13 did not form on any of the previously
used catalysts [4–6]; compounds 11 and 13 are not described
in the literature. For structure elucidation of compounds 11–13
weperformedtransformationoflimonenediepoxideswithlarger
amounts of the initial product. Dried methylene chloride (20 ml)
was added to sample 7 (2.00 g). Limonene diepoxides (1.50 g)
were added with stirring to the resulting suspension. After 5 h,
the reaction mixture was filtered off from the solid catalyst
through an alumina layer (second degree of activity, diethyl ether
as eluent). The mass of the reaction mixture was 1.40 g. A series
Methyl-8-methylen-2-oxabicyclo[3.3.1]non-3-en (13). The H
NMR spectrum contained the following signals (δ): 1.55 (d,
C10H3, J10,3 1.5), 1.69 (d.d.d, H9syn, J9syn,9an 12.5, J9syn,5 2.5,
J9syn,1 2), 1.94 (d.d.d.d, H9an, J 12.5, J9an,1 4, J9an,6 3.5, J9an,5
2.5), 2.36 (d.d.d.d.d, H7a, J7a,7e 14.5, J7a,6a 13, J7a,6e 6, J7a,11
2, J7a,11 2), 4.48 (m, H1), 4.71 (d.d, H11, J11,7a 2, J 2), 4.77
ꢀ
ꢀ
(d.d, H11 , J11 ,7a 2, J 2), 6.32 (q, H3, J3,10 1.5). The centers of
ꢀ
the following proton multiplets were determined from the 2D
13C–1H correlated spectrum on direct constants: 2.13 (m, H5),
1.54 m and 1.78 m (2H6), 2.16 (m, H7e). 13C NMR, δ: 74.74
(d, C1), 138.36 (d, C3), 109.41 (s, C4), 30.44 (d, C5), 30.00 (t,
C6), 27.69 (t, C7), 149.23 (s, C8), 31.23 (t, C9), 16.95 (q, C10),
111.33 (t, C11).
References
[1] J.G. Smith, Synthesis (1984) 629.
[2] V.G. Drjuk, V.G. Kartsev, M.A. Vojtsehovskaja, Oxiranes—Synthesis and
Biological Activity, Bogorodsky Pechatnik, Moscow, 1999, p. 528 (Rus.).
[3] N.F. Salakhutdinov, V.A. Barkhash, Usp. Khim. 66 (1997) 376;
N.F. Salakhutdinov, V.A. Barkhash, Russ. Chem. Rev. 66 (1997) 343.

80
O.V. Salomatina et al. / Journal of Molecular Catalysis A: Chemical 269 (2007) 72–80
[4] L.A. Mukhamedova, F.G. Nasybullina, M.I. Kudryavtseva, Izv. Akad.
Nauk. SSSR, Ser. Khim. 28 (1979) 847.
[26] T. Yamaguchi, Appl. Catal. 61 (1990) 1.
[27] A. Clearfield, G.P.D. Serrette, A.H. Khazi-Syed, Catal. Today 20 (1994)
295.
[28] M. Benaissa, J.D. Santiesteban, G. Diaz, C.D. Chang, M. Jose-Yacaman,
J. Catal. 161 (1996) 694.
[29] C.R. Vera, J.M. Parera, J. Catal. 165 (1997) 254.
[30] A.F. Bedilo, K.J. Klabunde, J. Catal. 176 (1998) 448.
[31] V. Parvulescu, S. Coman, P. Grange, V.I. Parvulescu, Appl. Catal. A 176
(1999) 27.
[5] O.V. Salomatina, O.I. Yarovaya, D.V. Korchagina, Y.u.V. Gatilov, M.P.
Polovinka, V.A. Barkhash, Zh. Org. Khim. 42 (2006) 1333;
O.V. Salomatina, O.I. Yarovaya, D.V. Korchagina, Y.u.V. Gatilov, M.P.
Polovinka, V.A. Barkhash, Russ. J. Org. Chem. 42 (2006) 1313.
[6] O.V. Salomatina, O.I. Yarovaya, D.V. Korchagina, M.P. Polovinka, V.A.
Barkhash, Mendeleev Commun. (2005) 59.
[7] Y. Noma, H. Nishimura, Agric. Biol. Chem. 51 (1987) 1845.
[8] G.T. Maatooq, A.A. Gohar, J.J. Hoffmann, Pharmazie 57 (2002) 282.
[9] S. Hatakeyama, M. Kawamura, E. Shimanuki, S. Takano, Tetrahedron Lett.
33 (1992) 333.
[10] M. Hino, S. Kobayashi, K. Arata, J. Am. Chem. Soc. 101 (1979) 6439.
[11] N. Katada, J. Endo, K. Notsu, N. Yasunobu, N. Naito, M. Niwa, J. Phys.
Chem. 104 (2000) 10321.
[12] C. Morterra, G. Cerrato, G. Meligrana, M. Signoretto, F. Pina, G. Strukul,
Catal. Lett. 73 (2001) 113.
[13] C.B. Chaabene, L. Bergaoui, A. Ghorbel, J.F. Lambert, P. Grange, Appl.
Catal. A 252 (2003) 411.
[14] M.G. Cutrufello, U. Diebold, R.D. Gonzalez, Catal. Lett. 101 (2005) 5.
[15] A. Corma, A. Martinez, C. Martinez, J. Catal. 149 (1994) 52.
[16] D.L. Negelein, R. Lin, R.L. White, J. Appl. Polym. Sci. 67 (1998) 341.
[17] M. Nino, K. Arata, Chem. Lett. (1981) 1671.
[18] G.K. Chuah, S.H. Liu, S. Jaenicke, L.J. Harrison, J. Catal. 200 (2001) 352.
[19] S. Ardizzone, C.L. Bianchi, G. Cappelletti, J. Porta, J. Catal. 227 (2004)
470.
[20] R.S. Drago, N. Kob, J. Phys. Chem. B 101 (1997) 3360.
[21] V. Adeeva, H.-Y. Liu, B.-Q. Xu, W.M.H. Sachtler, Top. Catal. 6 (1998) 61.
[22] D. Fa˜rcas¸iu, J.Q. Li, A. Kogelbauer, J. Mol. Catal. 124 (1997) 67.
[23] T.-K. Cheung, B.C. Gates, Top. Catal. 6 (1998) 41.
[24] C. Morterra, G. Cerrato, S. Ardizzone, C.L. Bianchi, M. Signoretto, F. Pina,
Phys. Chem. Chem. Phys. 4 (2002) 3136.
[32] S.X. Song, R.A. Kydd, J. Chem. Soc., Faraday Trans. 94 (1998)
1333.
[33] F. Babou, B. Bigot, P. Sautet, J. Phys. Chem. 97 (1993) 11501.
[34] F. Haase, J. Sauer, J. Am. Chem. Soc. 120 (1998) 13503.
[35] B. Umansky, J. Engelharrdt, W.K. Hall, J. Catal. 127 (1991) 128.
[36] K.T. Jung, A.T. Bell, J. Mol. Catal. A: Chem. 163 (2000) 27.
[37] T.G. Kuznetsova, V.A. Sadykov, E.M. Moroz, S.N. Trukhan, E.A. Pauk-
shtis, V.N. Kolomiichuk, E.B. Burgina, V.I. Zaikovskii, M.A. Fedotov, V.V.
Lunin, E. Kemnitz, Stud. Surf. Sci. Catal. 143 (2002) 659.
[38] V.A. Sadykov, T.G. Kuznetsova, V.P. Doronin, et al., Top. Catal. 32 (2005)
29.
[39] T. Kuznetsova, V. Sadykov, L. Batuev, E. Burgina, E. Moroz, et al., Proc.
Fourth Asia Pacific Congress on Catalysis, Singapore, 2006.
[40] V. Sadykov, T. Kuznetsova, V. Doronin, et al., Catal. Today 114 (2006)
13.
[41] E.P. Parry, J. Catal. 2 (1963) 371.
[42] D. Efremov, T. Kuznetsova, V. Doronin, V. Sadykov, J. Phys. Chem. B 109
(2005) 7451.
[43] E.F. Lopez, V.S. Escribano, M. Panizza, M.M. Carnasciali, G. Busca, J.
Mater. Chem. 11 (2001) 1891.
[44] V.A. Startzeva, L.A. Nikitina, V.V. Plemenkov, Zh. Org. Khim. 37 (2001)
46;
V.A. Startzeva, L.A. Nikitina, V.V. Plemenkov, Russ. J. Org. Chem. 37
(2001) 34.
[25] M. Waquif, J. Bachelier, O. saur, J.-C. lavalley, J. Mol. Catal. 72 (1992)
127.