Heteropoly acid catalysts for the synthesis of fragrance compounds from biorenewables: Isomerization of limonene oxide

Article abstract of DOI:10.1039/c2cy20526b

The liquid-phase isomerization of limonene oxide was studied in the presence of heteropoly acid catalysts in aprotic solvents in homogeneous and heterogeneous systems. Among the catalysts were bulk and silica-supported tungstophosphoric acid H₃PW₁₂O₄₀ and its acidic Cs salt Cs_0.5H_0.5PW₁₂O₄₀ (CsPW). The reaction gave dihydrocarvone, a valuable fragrance intermediate, as the main product with turnover numbers of up to 8000. The nature of the solvent had a strong effect on reaction rate and selectivity. CsPW (0.1 mol%) was found to be a highly efficient and truly heterogeneous catalyst for this reaction, providing 82% yield of dihydrocarvone in 1,4-dioxane as a solvent under ambient conditions. This simple catalytic method represents economically attractive route to industrially important compounds starting from bio-renewable substrates easily available from essential oils. The Royal Society of Chemistry 2013.

Full text of DOI:10.1039/c2cy20526b

View Online / Journal Homepage
Dynamic Article Links
Catalysis
Science & Technology
Cite this: DOI: 10.1039/c2cy20526b
www.rsc.org/catalysis
PAPER
Heteropoly acid catalysts for the synthesis of fragrance compounds from
biorenewables: isomerization of limonene oxidew
Vinı
´
cius V. Costa,a Kelly A. da Silva Rocha,b Ivan V. Kozhevnikov,c
Elena F. Kozhevnikovac and Elena V. Gusevskaya*a
Received 26th July 2012, Accepted 1st September 2012
DOI: 10.1039/c2cy20526b
The liquid-phase isomerization of limonene oxide was studied in the presence of heteropoly acid
catalysts in aprotic solvents in homogeneous and heterogeneous systems. Among the catalysts
were bulk and silica-supported tungstophosphoric acid H3PW12O40 and its acidic Cs salt
Cs0.5H0.5PW12O40 (CsPW). The reaction gave dihydrocarvone, a valuable fragrance intermediate,
as the main product with turnover numbers of up to 8000. The nature of the solvent had a strong
effect on reaction rate and selectivity. CsPW (0.1 mol%) was found to be a highly efficient and
truly heterogeneous catalyst for this reaction, providing 82% yield of dihydrocarvone in
1,4-dioxane as a solvent under ambient conditions. This simple catalytic method represents
economically attractive route to industrially important compounds starting from bio-renewable
substrates easily available from essential oils.
allylic alcohols: carveol and exo-carveol.4–10 For example, gold
nanoparticles supported on TiO2 are efficient catalysts for
selective isomerization of limonene oxide into exo-carveol.9
On the other hand, there are only a few reports describing selective
isomerization of limonene oxide into carbonyl compounds, such
as carvenone and dihydrocarvone.3,11–13 Dihydrocarvone has been
synthesized from limonene oxide in 50–70% yield using ZnBr2
as a homogeneous catalyst in benzene3 or solid LiClO4,11
BF3-etherate,11 or alumina6 in toluene, with all these catalysts
used in nearly stoichiometric amounts.
1. Introduction
Monoterpenic compounds are the main components of various
essential oils and represent an important renewable hydrocarbon
feedstock for the flavor and fragrance industry.1,2 Acid-catalyzed
transformations are used for the commercial production of
various terpenic fragrance compounds from more abundant
terpenoids. Many of these processes still employ mineral acids
as homogeneous catalysts, which results in serious technological
and environmental problems. The development of efficient solid
acid catalysts for the reactions involving terpenic compounds is
an important task for the valorization of original essential oils.
For example, limonene oxide, obtained from limonene
epoxidation or extracted from citric essential oils, can be
converted by acid-catalyzed transformations to various valuable
synthetic intermediates and ingredients for the flavor industry,
such as carveol, exo-carveol, dihydrocarvone and carvenone.3 This
reaction, however, usually gives a complex mixture of products
resulting from its isomerization, which can be accompanied by
addition of solvent or residual water. For this reason, the
development of selective syntheses of particular products from
limonene oxide is a challenge. In most reported procedures, the
isomerization of limonene oxide gives mainly para-menthenic
Heteropoly acids (HPAs) of the Keggin series have attracted
interest as promising acid catalysts for the clean synthesis of
fine and specialty chemicals.14,15 Due to their stronger acidity,
HPAs are usually more efficient catalysts compared to conventional
acid catalysts, such as mineral acids, zeolites, mixed oxides, and
ion-exchange resins. HPAs are less corrosive than mineral acids
and do not effect undesirable side reactions, such as sulfonation
and chlorination. HPAs are insoluble in non-polar solvents and
scarcely soluble in dry carboxylic acids, allowing truly hetero-
geneous acid catalysis in such media. On the other hand, in
many polar solvents, where HPAs are highly soluble, homo-
geneous catalysis is the only option. HPA salts with large
monovalent cations are insoluble in water and polar solvents
and can be prepared with large surface areas.16–18 In particular,
the hydrophobic acidic salt, Cs2.5H0.5PW12O40, which has strong
Brønsted acid sites and large surface area (100–150 m2 g–1), is a
very efficient solid-acid catalyst for a variety of liquid-phase
organic reactions.19–22
a
´
Departamento de Quımica, Universidade Federal de Minas Gerais
31270-901, Belo Horizonte, MG, Brazil. E-mail: elena@ufmg.br;
Fax: (+)55 31 34095700; Tel: (+)55 31 34095741
b
´
Departamento de Quımica, Universidade Federal de Ouro Preto,
35400-000, Ouro Preto, MG, Brazil
c Department of Chemistry, University of Liverpool,
Liverpool L69 7ZD, UK
In our previous studies, HPAs have been used as catalysts
for isomerization,23–26 hydration and esterification27–30 of terpenic
compounds in liquid phase. In a-pinene oxide isomerization23,24
w Electronic supplementary information (ESI) available. See DOI:
10.1039/c2cy20526b
c
This journal is The Royal Society of Chemistry 2012
Catal. Sci. Technol.

View Online
remarkable effect of solvent polarity and basicity on reaction
chemoselectivity has been found. This allows selective synthesis
of various valuable fragrance compounds through the choice of
appropriate solvent. Further using cerium and tin catalysts we
have shown that the balance between different reaction pathways
is mostly determined by the nature of solvent rather than by the
catalyst or other reaction variables.31
with stirring. The precipitate obtained was aged in aqueous
mixture for 48 h at room temperature and dried in a rotary
evaporator at 45 1C per 3 kPa and after that in an oven at 150 1C
per 0.1 kPa for 1.5 h. CsPW thus prepared had a surface area of
111 m2 gÀ1, pore volume 0.07 cm3 gÀ1, and pore diameter 24 A.
The acid strength of CsPW and silica-supported PW was
characterized calorimetrically using ammonia and pyridine
adsorption and discussed previously.35,36
Within our program aimed at adding value to natural
ingredients of renewable essential oils, we now investigate
the application of HPA catalysts for the isomerization of
limonene oxide, the reaction which has attracted much less
attention as compared to the widely studied isomerization of
a-pinene oxide. We explore heterogeneous HPA catalysis for
limonene oxide isomerization using Cs2.5H0.5[PW12O40]
(CsPW) and silica-supported H3PW12O40 as solid acid catalysts
in liquid phase in a range of aprotic solvents. Our primary goal
is to improve the selectivity to dihydrocarvone through catalyst
and solvent optimization. To our knowledge, no attempt to use
HPA catalysts for this reaction has been made so far.
2.4. Catalytic reactions
The reactions were carried out in a glass reactor equipped with
a magnetic stirrer and a condenser. In a typical run, a mixture
(total volume of 10.0 mL) of limonene oxide (1.5–3.0 mmol),
dodecane (1.0 mmol, GC internal standard) and the catalyst
(PW (3.5–17.5 mmol), PW/SiO2 (5 mg, 0.35 mmol of PW), or
CsPW (5 mg, 1.5 mmol)) in a specified solvent was intensely
stirred under air at a specified temperature (25–140 1C). The
reactions were followed by gas chromatography (GC) using a
Shimadzu 17 instrument fitted with a Carbowax 20 M capillary
column and a flame ionization detector. At indicated reaction
times, stirring was stopped and after quick catalyst settling
down aliquots were taken and analyzed by CG. The mass
balance, product selectivity and yield were determined using
dodecane as an internal standard.
2. Experimental methods
2.1. Chemicals
All chemicals were purchased from commercial sources and used as
received, unless otherwise stated. (+)-Limonene oxide (a mixture
of cis and trans isomers) was from Aldrich, H3PW12O40 hydrate
from Aldrich and Aerosil 300 silica from Degussa.
To control catalyst leaching and the possibility of a homo-
geneous reaction, the catalyst was removed by centrifugation
of the reaction mixture at the reaction temperature to avoid
re-adsorption of active components onto silica, then the super-
natant was added with a fresh portion of substrate, if necessary,
and allowed to react on. The absence of further reaction in such
experiments indicated the absence of PW leaching and vice
versa.
2.2. Characterization techniques
31P MAS NMR spectra were recorded at room temperature
and 4 kHz spinning rate on a Bruker Avance DSX 400 NMR
spectrometer using 85% H3PO4 as a reference. Powder X-ray
diffraction (XRD) of the catalysts was performed on a Rigaku
Geigerflex-3034 diffractometer with CuKa radiation. The textural
characteristics were determined from nitrogen physisorption
measured on a Micromeritics ASAP 2010 instrument at 77 K.
Tungsten and phosphorus content in HPA catalysts was deter-
mined by inductively coupled plasma (ICP atomic emission
spectroscopy) on a Spectro Ciros CCD spectrometer.
Products 2, 3 and 4 were separated by a column chromato-
graphy (silica gel 60) using mixtures of hexane and CH2Cl2 as
eluents and identified by GC-MS, 1H, and 13C-NMR. The 1H
and 13C-NMR signals were assigned using bidimensional
techniques. NMR spectra were recorded in CDCl3 using a Bruker
400 MHz spectrometer, with TMS as an internal standard. Mass
spectra were obtained on a Shimadzu QP2010-PLUS instrument
operating at 70 eV. Products 5 and 6 were identified by GC/MS
(Shimadzu QP2010-PLUS instrument, 70 eV) by comparison
with an authentic sample.
2.3. Catalyst preparation and characterization
The 20 wt% H3PW12O40/SiO2 catalyst (PW/SiO2) was prepared
by impregnating Aerosil 300 (SBET, 300 m2 gÀ1) with an aqueous
PW solution followed by calcination at 130 1C/0.2–0.3 Torr for
1.5 h, as described elsewhere.32 The PW content determined by
ICP was 20 wt%. The specific surface area determined by the
BET method was 200 m2 gÀ1. The average pore diameter and the
p-Mentha-8-ene-2-one (cis or trans) (2a, shorter GC retention
time) (dihydrocarvone). MS (EI, 70 eV): m/z 67 (100%), 95
(85%), 68 (55%), 109 (50%), 108 (50%), 82 (50%), 81 (50%),
69 (35%), 55 (30%) 110 (25%), 152 (M+, 20%), 137
1
(M+–CH3, 20%). H NMR (CDCl3, 400 MHz): d 4.76 (br s,
single point total pore volume were 144 A and 0.53 cm3 gÀ1
,
1H, C9HH), 4.73 (br s, 1H, C9HH), 2.40–2.45 (m, 2H, C3HH,
C4H), 2.25–2.35 (m, 1H, C3HH), 2.10–2.20 (m, 1H, C6HH),
1.85–1.95 (m, 1H, C5HH), 1.74 (s, 3H, C10H3), 1.55–1.65 (m,
respectively. The 31P MAS NMR spectrum of the catalyst
(Fig. S1, ESIw) displayed only a single peak at ca. À15 ppm,
characteristic of PW confirming the integrity of the Keggin
structure.33 From XRD (Fig. S2, ESIw), the catalyst included
finely dispersed PW on the silica surface, as practically no PW
crystal phase can be seen.
3
1H, C5HH), 1.35–1.40 (m, 1H, C6HH), 1.03 (d, J = 6.4 Hz,
3H, C7H3). 13C{1H} NMR (CDCl3, 100 MHz): d 212.55(C2),
147.63 (C8), 109.62 (C9), 47.04 (C4), 46.89 (C3), 44.75 (C1),
34.94 (C6), 30.80 (C5), 20.48 (C10), 14.27 (C7). (Fig. S3 and S4,
37
The acidic heteropoly salt CsPW was prepared according to
the literature procedure34 by adding dropwise the required
amount of aqueous solution of cesium carbonate (0.47 M) to
aqueous solution of H3PW12O40 (0.75 M) at room temperature
+
ESIw). Compound described Eros et al.
p-Mentha-8-ene-2-one (trans or cis) (2b) (dihydrocarvone).
MS (EI, 70 eV): m/z 67 (100%), 95 (95%), 68 (55%), 82 (50%),
c
Catal. Sci. Technol.
This journal is The Royal Society of Chemistry 2012

View Online
69 (40%), 81 (35%), 55 (32%), 152 (M+, 30%), 109 (30%),
137 (M+–CH3, 20%), 110 (20%). 1H NMR (CDCl3,
400 MHz): d 4.83 (br s, 1H, C9HH), 4.69 (br s, 1H, C9HH),
2.55–2.65 (m, 1H, C4H), 2.50–2.55 (m, 1H, C3HH), 2.35–2.45
(m, 1H, C3HH), 1.80–1.90 (m, 3H, C5H2, C6HH), 1.74 (s, 3H,
p-Mentha-1(7),8-diene-2-ol (5) (exo-carveol). MS (EI, 70 eV):
m/z 109 (100%), 55 (75%), 67 (70%), 69 (65%), 91 (58%), 119
(55%), 79 (55%), 93 (50%), 95 (46%), 81 (40%), 83 (35%), 84
(30%), 134 (M+–H2O, 25%), 105 (25%), 77 (25%), 152 (M+,
5%). (Fig. S3, ESIw).
3
C10H3), 1.55–1.65 (m, 1H, C6HH), 1.09 (d, J = 7.2 Hz, 3H,
C7H3). 13C{1H} NMR (CDCl3, 100 MHz): d 212.03 (C2),
Carveol (6). MS (EI, 70 eV): m/z 109 (100%), 84 (59%), 67
(40%), 91 (32%), 95 (30%), 83 (30%), 55 (30%), 69 (28%),
152 (M+, 6%). (Fig. S3, ESIw). For NMR see ref. 23.
146.89 (C8), 111.57 (C9), 44.65 (C1), 44.16 (C3), 43.94 (C4),
30.65 (C6), 26.38 (C5), 21.61 (C10), 15.61 (C7). (Fig. S3 and S5,
37
+
ESIw). Compound described Eros et al.
3. Results and discussion
1-Methyl-3-isopropenylcyclopentyl-1-carboxaldehyde (3). MS
(EI, 70 eV): m/z 81 (100%), 137 (M+–CH3, 75%), 67 (75%), 55
(60%), 82 (45%), 93 (40%), 79 (40%), 109 (38%), 71 (35%),
123 (32%), 95 (29%), 53 (27%), 69 (24%), 91 (19%), 119
The results of limonene oxide (1) isomerization in the presence
of PW catalysts in various solvents are presented in Table 1. In
all solvents, limonene oxide was stable in the absence of
catalysts. In blank reactions (not shown in Table 1), only
negligible conversions were observed in 6 h. Detected in
reaction were the following products (Scheme 1): dihydrocarvone
(2), 1-methyl-3-isopropenyl-cyclopentyl-1-carboxaldehyde (3),
limonene-1,2-diol (4), exo-carveol (5) and carveol (6). Our
efforts were directed to achieve high selectivity to any of these
products, especially 2, through the choice of catalyst and
solvent and to perform the process under truly heterogeneous
conditions, i.e., in the absence of leaching of HPA from the
catalyst. As PW is insoluble in cyclohexane, dichloromethane
and dichloroethane, we started with PW/SiO2 catalyst using
these solvents.
1
(16%), 152 (M+, 15%). H NMR (CDCl3, 400 MHz): d 9.43
(s, 1H, C6H), 4.64 (br s, 1H, C9HH), 4.63 (br s, 1H, C9HH),
2.30–2.40 (m, 1H, C3H), 2.10–2.20 (m, 1H, C2HH), 1.90–2.00
(m, 1H, C5HH), 1.70–1.80 (m, 1H, C4HH), 1.65 (s, 3H, C10H3),
1.30–1.40 (m, 1H, C5HH), 1.20–1.25 (m, 1H, C4HH), 1.15–1.25
(m, 1H, C2HH), 1.11 (s, 3H, C7H3). 13C{1H} NMR (CDCl3,
100 MHz): d 203.51 (C6), 146.32 (C8), 107.81 (C9), 52.24 (C1),
45.95 (C3), 38.70 (C2), 32.19 (C5), 29.50 (C4), 21.11 (C10), 20.17
(C7). (Fig. S3 and S6, ESIw).
p-Mentha-8-ene-1,2-diol (cis) (4) (limonene 1,2-diol, cis). MS
(EI, 70 eV): m/z 71 (100%), 72 (56%), 108 (50%), 73 (39%),
109 (36%), 93 (33%), 82 (32%), 137 (M+–H2O–CH3, 28%),
58 (28%), 55 (27%), 152 (M+–H2O, 22%), 170 (M+, 1%).
1H NMR (CDCl3, 400 MHz): d 4.70 (br s, 2H, C9H2), 3.69
(br s, 1H, C2H), 2.20–2.30 (m, 1H, C4H), 1.85–1.90 (m, 1H,
C3HH), 1.74 (s, 3H, C10H3), 1.70–1.80 (m, 1H, C6HH),
1.60–1.70 (m, 1H, C3HH), 1.45–1.60 (m, 3H, C5H2, C6HH),
1.27 (s, 3H, C7H3). 13C{1H} NMR (CDCl3, 100 MHz): d
149.21 (C8), 108.96 (C9), 73.76 (C2), 71.42 (C1), 37,43 (C4),
34.46 (C3), 34.01 (C6), 26.39 (C7), 26.10 (C5), 21.05 (C10).
(Fig. S3 and S7, ESIw).
It should be mentioned that the total selectivity for the
detected products in some runs did not reach 100% mainly
due to the formation of high-boiling products, which were not
observable by GC. In the presence of only 0.1 wt% of PW/
SiO2 in cyclohexane (relative to the whole reaction mixture), a
90% conversion of limonene oxide was observed in 10 min at
25 1C, with the reaction nearly completed in 1 h (Table 1, run 1).
Dihydrocarvone was detected as the main product formed with
47% selectivity as a mixture of exo and endo isomers in
comparable amounts. Along with dihydrocarvone, diol 4 and
Table 1 Isomerization of limonene oxide catalyzed by H3PW12O40 (PW)a
Product selectivity (%)
Run Solvent
Catalyst (mg) PW/mmol T/1C Time/min Conversion (%)
2
3
4
5
6
TONb TOFb (minÀ1
385
)
1c
Cyclohexane
PW/SiO2 (5)
0.35
25
10
60
90
97
42
47
7
5
19 tr. tr.
14 tr. tr. 4156
2c
3c
4d
5
Dichloromethane PW/SiO2 (5)
Dichloroethane PW/SiO2 (5)
Dichloromethane PW/SiO2 (5)
0.35
0.35
0.35
3.5
25
25
25
25
10
10
90
10
100
100
96
44 12 15 tr. tr. 4285
48 23 17 tr. tr. 4285
58 22
65
65
36 12 22
tr.
62 tr. tr. 19 15
63 tr. tr. 16 15
56
56
50
53
428
428
91
9
14
12
tr. tr. 8228
1,4-Dioxane
PW (10)
62
5
8
1
4
2
2
7
7
7
60
15
100
100
o5
36
70
85
93
93
96
428
428
6
7
8
9
10
PhNO2
DMA
DMA
DMA
DMA
PW (10)
PW (10)
PW (10)
PW (10)
PW (10)
3.5
3.5
3.5
3.5
3.5
25
25
100
130
140
28
480
240
240
240
480
60
154
300
0.6
1.3
4
4
4
4
4
5
9
5
18 10
18 12
18 10
20 15
400
80
82
0.8
1.3
1.5
11
12
DMA
DMA
PW (50)
PW (50)
17.5
17.5
100
140
60
a
Substrate 1.5 mmol, total volume 10 mL. Conversion and selectivity were determined by GC; the difference in mass balance was due to the
b
formation of high-boiling products; tr. – trace amounts; DMA – dimethylacetamide. TON (turnover number) in moles of substrate converted per
c
mole of a total amount of PW in catalyst. TOF – the average turnover frequency. After runs 1, 2 and 3, the catalyst was removed, the solution
was recharged with fresh substrate (1.5 mmol) and the reactions were allowed to proceed further, with no further conversion observed thereupon.
d
Substrate 3.0 mmol.
c
This journal is The Royal Society of Chemistry 2012
Catal. Sci. Technol.

View Online
Scheme 1 Main products of acid-catalyzed transformations of limonene oxide (1).
ring contraction product 3 were formed in small amounts. High
Table 2 Dielectric constants (e) for the solvents and pKa values for
corresponding conjugated acids38
turnover number and turnover frequency per mol of the total
amount of PW were achieved in this run (TON = 4156 and
initial TOF = 385 minÀ1). But the total yield of isomerization
products 2 and 3 in cyclohexane was rather low (ca. 50%). The
formation of diol 4 (14%) was probably due to epoxide ring
opening by adventitious water present in the porous PW/SiO2
catalyst, commercial solvent and/or limonene oxide.
Approximate pKa
(relative to water)38
Solvent
e
Conjugated acid
DMA
PhNO2
Dichloroethane
Dichloromethane
1,4-Dioxane
Cyclohexane
37.8
34.8
10.4
8.9
2.2
2.0
DMAH+
PhNO2H+
—
—
C4H8O2H+
—
À0.5
À11
—
—
À3.5
Further it has been found that the nature of solvent exerts
remarkable effect on acid-catalyzed transformations of limo-
nene oxide. In more polar solvents, such as dichloroethane and
dichloromethane, much better reaction selectivities were
obtained than in cyclohexane. Isomerization products 2 and
3 were formed in ca. 70% total yield along with ca. 15% of
diol 4 (Table 1, runs 2 and 3). The reactions were very fast at
25 1C and showed complete conversion at first sampling in
10 min, with average TOFs of ca. 430 minÀ1. At a higher
substrate concentration, the selectivity to diol 4 decreased
probably due to a smaller water/substrate ratio, with the total
yield of 2 and 3 as high as 80% (Table 1, run 4).
—
highly polar and basic solvent (Table 1, runs 7–12). In DMA,
allylic alcohols 5 and 6 were formed in 35% total selectivity
together with 50–65% of dihydrocarvone. 90% total yield of 2,
5 and 6 was obtained under optimized conditions in DMA
(Table 1, runs 11 and 12). Aldehyde 3 and diol 4 were detected
in DMA only in small amounts.
At room temperature in the presence of PW, limonene oxide
was nonreactive in DMA (Table 1, run 7). But at 100 1C, fast
conversion of limonene oxide into products 2, 5 and 6 with a
total selectivity of 96% was observed (Table 1, run 8). The
reaction, however, stopped at about 35% conversion in 60 min
(Fig. 1). At higher temperature (140 1C), the reaction went
almost to completion (Fig. 1) to give dihydrocarvone with
60% selectivity and allylic alcohols 5 and 6 with 30% combined
selectivity (Table 1, run 10). This behavior was not the result of
thermodynamic control since addition of larger amounts of
catalyst increased the conversion, as illustrated by runs 8 and 11
(Table 1). Pre-treatment of DMA with Amberlyst-15 acidic resin
to remove possible adventitious basic impurities had no effect on
the reaction. Therefore a likely explanation is that the catalyst
was partially deactivated by interaction with the solvent.
The possibility of HPA leaching from the PW/SiO2 catalyst
and any contribution of homogeneous catalysis in reaction
in cyclohexane, dichloromethane and dichloroethane was verified
in special experiments. After runs 1, 2, and 3 (Table 1), the catalyst
was separated by centrifugation, and the filtrates were recharged
with fresh substrate and allowed to react. No further reaction was
observed after catalyst removal, which shows the absence of any
contribution of homogeneous catalysis, implying that there was no
significant HPA leaching in these reaction systems.
To examine the effect of solvent on limonene oxide isomer-
ization, we have tested several aprotic solvents which readily
dissolved PW and could affect homogeneous catalysis
(Table 1, runs 5–12). These solvents included 1,4-dioxane,
nitrobenzene and dimethylacetamide (DMA), which differ in
their polarity and basicity (Table 2). The polarity is represented
by dielectric constants and the basicity by pKa values for the
corresponding conjugate acids. In 1,4-dioxane, a non-polar
basic solvent, dihydrocarvone, was obtained in 65% yield along
with small amounts of 3, 5, and 6 at 25 1C (Table 1, run 5). It is
evident that the reaction in rather basic dioxane is much slower
than in neutral cyclohexane, dichloromethane and dichloro-
ethane, regardless of the solvent polarity (Table 1, compare
runs 1, 2 and 3 with run 5 with 10 times more PW). On the other
hand, in highly polar but less basic nitrobenzene, the reaction
was faster than in 1,4-dioxane, although less selective to
dihydrocarvone (Table 1, run 6 vs. run 5).
The effect of solvent on reaction selectivity was especially
pronounced when the reaction was performed in DMA, a
Fig. 1 Isomerization of limonene oxide (0.15 M) catalyzed by
H3PW12O40 (0.35 mM) in DMA at different temperatures.
c
Catal. Sci. Technol.
This journal is The Royal Society of Chemistry 2012

View Online
Scheme 2 Schematic representation of acid-catalyzed transformations of limonene oxide (1).
At 140 1C, even small amounts of PW converted all substrate
formed in appreciable amounts. Finally, A can rearrange into
in 8 h with a TON of 400 (Table 1, run 10). Thus, the
deactivation is probably reversible since the catalyst can be
re-activated by increasing the temperature. Similar effect has
been observed previously in isomerization of a-pinene oxide in
DMA.7 In this regard, high boiling point of DMA (166 1C) is
an important advantage for these reactions.
carbenium ion C through electron pair transfer from C2–C3
s-bond to C3. Then, C can lose a proton to give the ring
contraction product, aldehyde 3.
Thus, the application of PW catalysts for the isomerization
of limonene oxide allows obtaining dihydrocarvone in a
50–60% yield along with only one or two minor products:
aldehyde 3 or carveol and exo-carveol, which are also valuable
fragrance compounds. The best selectivities were obtained
under homogeneous rather than heterogeneous conditions in
oxygenated solvents (1,4-dioxane and DMA) which dissolve
PW. The use of solid PW/SiO2 catalyst in these solvents is not
possible due to PW leaching.
In Table 1 one can see that TON and TOF values are about
one order of magnitude lower for PW in homogeneous systems
than for the supported PW catalyst in heterogeneous systems.
This can be explained by the solvent effect on the strength
of active proton sites. Protons in the homogeneous systems
(1,4-dioxane, DMA and PhNO2) are strongly solvated hence
weaker and less active than those much less solvated in the
heterogeneous systems (cyclohexane, dichloromethane and
dichloroethane). The TON and TOF values were calculated
per total number of PW molecules in the catalysts, assuming
that all the protons were active. This is obviously true
for homogeneous PW systems and also appears to be true
for PW/SiO2 at moderate PW loadings (r20%).36
Further, we tested the CsPW salt as the catalyst for limonene
oxide isomerization. Importantly, the lack of solubility of CsPW as
compared to PW allowed the development of truly heterogeneous
synthesis of dihydrocarvone in appropriate solvents. Representa-
tive results for the reaction at 25 1C are given in Table 3. The
reaction in cyclohexane was very fast in the presence of small
amounts of CsPW (0.1 mol%); it reached 90% in the first aliquot
taken for analysis in 10 min. Dihydrocarvone was obtained in
higher yield (60%), than with PW/SiO2 under similar conditions
(Table 1, run 1 vs. Table 3, run 1). The advantage of CsPW over
PW in terms of reaction selectivity was even greater in other
solvents. The reactions in dichloromethane, dichloroethane and
nitrobenzene gave dihydrocarvone and aldehyde 3 in nearly 70 and
20% yields, respectively (Table 3, runs 2, 3, 4, and 7). These results
correspond to TONs of 1000–2000 per mol of the total amount of
CsPW. Considering that a part of acid sites may be located in the
bulk of the solid phase and hence not accessible to the substrate,
the real efficiency of the surface active sites could be even higher.
The proposed mechanism for the acid-catalyzed transfor-
mations of limonene oxide 1 into products 2–6 is shown in
Scheme 2. The opening of epoxy ring in 1 is induced by
protonation of the oxygen atom to give carbenium ion A.
The latter can undergo several competing transformations. It
can form dihydrocarvone 2 through C2–C1 hydride shift
followed by proton abstraction from carbenium ion B. Attack
of water on A can result in the formation of limonene diol 4.
Deprotonation of A at C7 or C6 can give directly exo-carveol 5
or carveol 6. Deprotonation could be assisted by basic solvents,
such as DMA and dioxane, in which products 5 and 6 were
c
This journal is The Royal Society of Chemistry 2012
Catal. Sci. Technol.

View Online
Table 3 Isomerization of limonene oxide catalyzed by Cs2.5H0.5PW12O40 (CsPW)a
Product selectivity (%)
Run
1c
Solvent
Time (min)
Conversion (%)
2
3
4
5
6
TONb
TOFb (minÀ1
)
Cyclohexane
10
60
240
240
360
10
60
10
60
15
90
98
100
100
96
90
100
91
50
63
69
68
72
81
82
80
80
71
12
15
20
16
20
6
7
9
6
17
25
10
8
10
6
4
4
5
6
5
4
tr.
tr.
tr.
1
2
3
2
tr.
3
2
tr.
tr.
tr.
1
2
3
3
tr.
980
1000
1000
1920
16.3
4.2
4.2
5.3
2c
3b
4d
5c
Dichloromethane
Dichloroethane
Dichloromethane
1,4-Dioxane
1000
16.7
6d
1,4-Dioxane
PhNO2
98
98
1960
980
32.7
65.3
7c
6
a
Substrate 1.5 mmol, catalyst 5 mg (1.5 mmol); total volume 10 mL, 25 1C. Conversion and selectivity were determined by GC; the difference in
mass balance was due to the formation of high-boiling products; tr. – trace amounts; DMA – dimethylacetamide. TON in moles of substrate
b
c
converted per mole of a total amount of CsPW, TOF – the average turnover frequency. After runs 1, 2, 5 and 7, the catalyst was removed, the
solution was recharged with fresh substrate (1.5 mmol) and the reactions were allowed to proceed further, with no further conversion observed
d
thereupon. Substrate 3.0 mmol.
Although dihydrocarvone was the major reaction
product in all solvents tested, the highest dihydrocarvone yield
(82%) was obtained in 1,4-dioxane (Table 3, runs 5 and 6).
This is probably the best result ever reported for this
reaction. CsPW is a highly efficient catalyst, operating
under ambient conditions. It is truly heterogeneous, as con-
firmed by the filtration test described above, and can be
separated from the reaction mixture by simple centrifugation
and reused.
valuable compounds starting from bio-renewable substrates
easily available from essential oils.
Acknowledgements
Financial support and scholarships from CNPq, CAPES,
FAPEMIG, and INCT-Catalise (Brazil) are acknowledged.
The nature of aprotic solvent used was found to affect
significantly both the reaction rate and selectivity. The effect
of solvent was similar for both catalysts studied and can be
summarized as follows. The solvent polarity (dielectric
constant) does not seem to have strong effect on the reaction
rate (conversion). Non-polar solvents, such as cyclohexane,
dichloromethane and dichloroethane, and polar nitrobenzene
showed high reaction rates with PW and CsPW catalysts. In
contrast, the solvent basicity had profound influence on both
the rate and selectivity of limonene oxide isomerization. The
more basic solvents, DMA and 1,4-dioxane, tend to predictably
decrease the reaction rate due to decreasing catalyst acid
strength. At the same time, these solvents show higher selectiv-
ities to dihydrocarvone. This may be explained assuming that in
the basic ‘‘slow’’ solvents the reaction occurs under kinetic
control, whereas in the non-basic ‘‘fast’’ solvents the
reaction is equilibrium controlled. Therefore by choosing the
appropriate solvent the reaction can be directed towards the
desired product(s).
Notes and references
1 W. E. Erman, Chemistry of Monoterpenes. An Encyclopedic Handbook,
Marcel Dekker, New York, 1985.
2 C. Sell, in The Chemistry of Fragrances: from Perfumer to
Consumer, ed. C. Sell, RSC Publishing, Dorset, UK, 2nd edn,
2006, vol. 2, pp. 52–88.
3 R. L. Settine, G. L. Parks and G. L. K. Hunter, J. Org. Chem.,
1964, 29, 616–618.
4 I. C. Nigam and L. Levi, Can. J. Chem., 1968, 46, 1944–1947.
5 E. H. Eschinasi, Isr. J. Chem., 1968, 6, 713–721.
6 K. Arata and K. Tanabe, Chem. Lett., 1976, 321–322.
7 J. Jayasree and C. S. Narayanan, Res. Chem. Intermed., 1997, 23,
169–178.
8 EP Pat., 1 404 635, 2003.
9 C. Raptis, H. Garcia and M. Stratakis, Angew. Chem., Int. Ed.,
2009, 48, 3133–3136.
10 US Pat., 7 884 252, 2011.
11 K. Arata, S. Akutagatva and K. Tanabe, J. Catal., 1976, 41, 173–179.
12 JP Pat., 62114926, 1987.
13 T. Kurata, T. Koshiyama and H. Kawashima, Yukagaku, 1985, 34,
1032–1034.
14 I. V. Kozhevnikov, Catalysts for Fine Chemicals, Catalysis by
Polyoxometalates, Wiley, Chichester, 2002, vol. 2.
15 T. Okuhara, N. Mizuno and M. Misono, Appl. Catal., A, 2001,
222, 63–77.
16 I. V. Kozhevnikov, Appl. Catal., A, 2003, 256, 3–18.
17 J. B. Moffat, Metal-Oxygen Clusters. The Surface and Catalytic
Properties of Heteropoly Oxometalates, Kluwer Academic Publishers,
New York, 2001.
18 H. Niiyama, Y. Saito, S. Yoshida and E. Echigoya, Nippon Kagaku
Kaishi, 1982, 4, 569–573.
19 J. B. Moffat, J. Mol. Catal., 1989, 52, 169–191.
20 T. Okuhara, N. Mizuno and M. Misono, Adv. Catal., 1996, 41, 113–252.
21 M. Misono and T. Okuhara, CHEMTECH, 1993, 23, 23–29.
22 T. Okuhara, T. Nishimura, H. Watanabe and M. Misono, Stud.
Surf. Sci. Catal., 1994, 90, 419–428.
4. Conclusions
Silica-supported PW and its acidic Cs salt are efficient, environ-
mentally benign and versatile heterogeneous catalysts for the
liquid-phase isomerization of limonene oxide. The reaction can
be performed under ambient conditions at low catalyst loadings
and gives dihydrocarvone as the main product. High yields of
this valuable fragrance intermediate were achieved through the
choice of the solvent, whose polarity and basicity strongly
affected the reaction rate and selectivity. This simple catalytic
method represents economically attractive route to industrially
23 K. A. da Silva Rocha, I. V. Kozhevnikov and E. V. Gusevskaya,
Appl. Catal., A, 2005, 294, 106–110.
c
Catal. Sci. Technol.
This journal is The Royal Society of Chemistry 2012

View Online
24 K. A. da Silva Rocha, J. L. Hoehne and E. V. Gusevskaya,
Chem.–Eur. J., 2008, 14, 6166–6172.
25 K. A. da Silva Rocha, P. A. Robles-Dutenhefner, E. M. B. Sousa,
E. F. Kozhevnikova, I. V. Kozhevnikov and E. V. Gusevskaya,
Appl. Catal., A, 2007, 317, 171–174.
26 K. A. da Silva Rocha, P. A. Robles-Dutenhefner, I. V. Kozhevnikov
and E. V. Gusevskaya, Appl. Catal., A, 2009, 352, 188–192.
27 P. A. Robles-Dutenhefner, K. A. da Silva, M. R. H. Siddiqui,
I. V. Kozhevnikov and E. V. Gusevskaya, J. Mol. Catal. A: Chem.,
2001, 175, 33–42.
31 V. V. Costa, K. A. da Silva Rocha, L. F. de Sousa, P. A. Robles-
Dutenhefner and E. V. Gusevskaya, J. Mol. Catal. A: Chem., 2011,
345, 69–74.
32 I. V. Kozhevnikov, A. Sinnema, A. J. A. van der Weerdt and
H. van Bekkum, J. Mol. Catal. A: Chem., 1997, 120, 63–70.
33 I. V. Kozhevnikov, Chem. Rev., 1998, 98, 171–198.
34 Y. Izumi, M. Ono, M. Kitagawa, M. Yoshida and K. Urabe,
Microporous Mater., 1995, 5, 255–262.
35 E. F. Kozhevnikova and I. V. Kozhevnikov, J. Catal., 2004, 224,
164–169.
28 K. A. da Silva, I. V. Kozhevnikov and E. V. Gusevskaya, J. Mol.
Catal. A: Chem., 2003, 192, 129–134.
36 A. M. Alsalme, P. V. Wiper, Y. Z. Khimyak, E. F. Kozhevnikova
and I. V. Kozhevnikov, J. Catal., 2010, 276, 181–189.
+
29 K. A. da Silva Rocha, N. V. S. Rodrigues, I. V. Kozhevnikov and
E. V. Gusevskaya, Appl. Catal., A, 2010, 374, 87–94.
30 A. L. P. de Meireles, K. A. da Silva Rocha, I. V. Kozhevnikov and
E. V. Gusevskaya, Appl. Catal., A, 2011, 409–410, 82–86.
37 G. Eros, H. Mehdi, I. Papai, T. A. Rokob, P. Kiraly, G. Tarkanyi
and T. Soos, Angew. Chem., Int. Ed., 2010, 49, 6559–6563.
38 J. March, Advanced Organic Chemistry: Reactions, Mechanisms,
and Structures, Wiley, NY, 4th edn, 1992, p. 250.
c
This journal is The Royal Society of Chemistry 2012
Catal. Sci. Technol.