One-pot synthesis of phenols from aromatic aldehydes by Baeyer-Villiger oxidation with H2O2 using water-tolerant Lewis acids in molecular sieves

Article abstract of DOI:10.1016/S0021-9517(03)00291-4

The heterogeneous oxidation systems Sn-Beta/H₂O₂ and Al-Beta/H₂O₂ were both active for the Baeyer-Villiger oxidation of aromatic aldehydes. With alkoxy substituents in ortho and especially in para position, the corresponding formate ester was the primary reaction product with excellent selectivity. The highest selectivity toward the ester was obtained with the Sn-Beta catalyst in dioxane as solvent. High selectivities of the corresponding phenol could be obtained by employing ethanol or aqueous acetonitrile as solvent with Sn-Beta. Al-Beta was more efficient as catalyst for both Baeyer-Villiger oxidation and ester hydrolysis and, thus achieved high yields for the corresponding alcohol. However, this was a less selective catalyst than Sn-Beta zeolite. Sn-Beta was a chemoselective catalyst when the aldehyde reactant contained olefinic groups in an alkyl chain. With these reactants, the Al-zeolite gave no activity. Other classical BV oxidants, e.g., peracids, also yielded the epoxidation products. Thus, Sn-Beta catalysts open a new entry to aromatic phenols with unsaturated alkyl substituents that are interesting intermediates in the chemical industry.

Full text of DOI:10.1016/S0021-9517(03)00291-4

Journal of Catalysis 221 (2004) 67–76
www.elsevier.com/locate/jcat
One-pot synthesis of phenols from aromatic aldehydes by Baeyer–Villiger
oxidation with H2O2 using water-tolerant Lewis acids in molecular sieves
Avelino Corma,∗ Vicente Fornés, Sara Iborra, María Mifsud, and Michael Renz
Instituto de Tecnología Química, UPV-CSIC, Universidad Politécnica de Valencia, Avda. de los Naranjos s/n, 46022 Valencia, Spain
Received 24 April 2003; revised 23 June 2003; accepted 23 June 2003
Abstract
Sn-Beta zeolite is employed as catalyst in the Baeyer–Villiger oxidation of aromatic aldehydes. Aldehydes involving alkoxy substituents
are oxidized to the corresponding formate ester which are hydrolyzed to the corresponding phenols. Choosing the adequate reaction condi-
tions the main product can be preselected. Dioxane solvent and a hydrogen peroxide deficit give predominantly the ester whereas ethanol
solvent or aqueous acetonitrile favor the phenol product. Double bonds in alkyl side chains do not react and the corresponding unsaturated
phenols are obtained with very high chemoselectivity. Brønsted sites are active for the Baeyer–Villiger oxidation of aromatic aldehydes with
H O provided that the molecule does not contain olefinic groups. In any case, the selectivity of Brønsted acid zeolites is lower than that of
2
2
Sn-Beta.
2003 Elsevier Inc. All rights reserved.
Keywords: Oxidation aromatic aldehydes; Baeyer–Villiger oxidation aldehydes; Baeyer–Villiger oxidation with hydrogen peroxide; Sn-Beta Lewis acid
catalyst
1. Introduction
from alkoxy benzene [10]. The Baeyer–Villiger oxidation
of alkoxybenzaldehydes has been extensively studied with
ortho-, meta-, and para-anisaldehyde (methoxybenzalde-
hyde) as model substrate. Several oxidants such as meta-
chloroperbenzoic acid (mCPBA) [10], monopersuccinic
acid [11], H2O2/seleninic acid [12], H2O2/methanol/H2SO4
[13], and H2O2/formic acid [14] have been tested for this re-
action. The percarboxylic and perseleninic acids gave good
to excellent yields of the corresponding phenol for the or-
tho- and the para-isomer [10–12]. The meta-anisaldehyde is
transformed into meta-anisic acid predominantly since the
methoxy group in meta position does not favor the migra-
tion of the aromatic moiety. The drawback of these oxidation
methods is the waste produced that in the case of the percar-
boxylic acids is at least one molecule of the acid, while in
the case of perseleninic acid considerable amounts of cata-
lyst are needed since within 30 h turnover numbers (TON)
as low as 12 have been reported [12]. The BV oxidation
with hydrogen peroxide catalyzed by sulfuric acid gives ex-
cellent yields of the desired products; however, water-free
conditions and highly concentrated hydrogen peroxide were
necessary since the reaction is believed to proceed via the
peroxy hemiacetal [13]. The oxidation with hydrogen per-
oxide in formic acid is less suitable for the production of
Substituted phenols, specially 4-alkoxy derivatives, are
important substances in organic chemistry since they are in-
termediates for drugs, agrochemicals, and dyes [1]. They
are useful as polymerization inhibitors for vinyl monomers
and stabilizers for polyesters [1] as well as antioxidants
for foods and cosmetics [2] and have been evaluated for
anti-lipid-peroxidation activity in rat liver microsomes [3].
Furthermore, they have shown physiological activity. para-
Alkoxyphenols can be employed for treating melanomas
[4,5] since they deactivate the ribonucleotide reductase in tu-
mor cells and, thereby, inhibit the growth of tumor cells [6].
This inhibition of the ribonucleotide reductase can also be
applied in AIDS therapy [7].
The 4-alkoxyphenols can be prepared by monoalkylation
of hydroquinone [3,4,8]. However, di-alkylation can also
occur decreasing the yield of the desired substrate. An alter-
native synthesis of the mono-alkoxyphenols starts with the
Baeyer–Villiger (BV) oxidation of aromatic aldehydes [9]
which are readily prepared by the Vilsmeier–Haack method
*
Corresponding author.
E-mail address: acorma@itq.upv.es (A. Corma).
0021-9517/$ – see front matter
2003 Elsevier Inc. All rights reserved.
doi:10.1016/S0021-9517(03)00291-4

68
A. Corma et al. / Journal of Catalysis 221 (2004) 67–76
1
phenols since it has been developed for the general oxida-
tion of aldehydes to carboxylic acids. There is then much
incentive to find new environmentally friendly catalysts and
processes for converting the substituted aromatic aldehydes
to the corresponding phenols.
Detector 5973 Network. H NMR spectra were recorded
with a Bruker spectrometer at a frequency of 300 MHz and
13C spectra at a frequency of 75 MHz.
2.4. Synthesis of 4-(3-methylbut-2-enoxy)-benzaldehyde
We have recently introduced Sn-Beta zeolite and Sn-
MCM-41 catalysts that are active and selective using aque-
ous H2O2 as oxidant while producing only H2O as subprod-
uct [15–17]. We will show here that by properly coupling
catalyst design and reaction conditions very high turnover
numbers and selectivity to phenols can be obtained in a one-
pot reaction system. It will also be shown that this is at
present an excellent alternative for producing phenolic mole-
cules with pharmaceutical interest when an unsaturated alkyl
group is present.
(1j) [19]
4-Hydroxybenzaldehyde (2.4 g, 20 mmol) and 3.27 g of
4-bromo-2-methyl-2-butene were dissolved in 60 mL dry
dimethylformamide. Potassium iodide (0.16 g, 1.5 mmol)
and 11 g (78 mmol) of fine powdered potassium carbon-
ate were added and the reaction mixture was stirred for
2 h at room temperature. Water (ca. 30 mL) was added and
the mixture extracted with methyl tert-butyl ether (4 times,
20 mL). The combined organic phases were washed with
30 mL of 10% HCl and with 30 mL of 10% aqueous K2CO3
solution. After drying with MgSO4, the solvent was removed
with a rotavaporator and 2.8 g (15 mmol, 74%) of a pale yel-
low liquid was obtained.
2. Experimental
2.1. Synthesis of the catalysts
MS m/z (%): 190 (7) [M]+, 121 (100), 93 (8), 77 (5), 69
(96), 53 (10), 41 (46). 1H NMR (300 MHz, CDCl3, 298 K):
δ = 1.76 (s, 3 H, 9-H or 10-H), 1.81 (s, 3 H, 9-H or 10-H),
4.63 (d, J = 12.5 Hz, 2 H, 6-H), 5.46 (m, 1 H, 7-H), 7.00
(d, J = 8.7 Hz, 2 H, 4-H), 7.83 (d, J = 8.7 Hz, 2 H, 3-H),
9.89 (s, 1 H, 1-H). 13C NMR (75 MHz, CDCl3, 298 K): δ =
194.4, 167.6, 142.7, 135.6, 122.4, 118.6, 68.8, 29.4, 21.9.
All Sn-Beta, Al-Beta, and Sn,Al-Beta zeolites with the
Sn and Al contents as stated in Table 4 were synthesized ac-
cording to the literature procedure [16,18]. Sn contents were
determined by chemical analysis. The Beta zeolites were cal-
cined at 853 K for 3 h. A high crystallinity of the zeolites was
observed by XRD, and in the case of Sn-Beta samples no
peaks of SnO2 were found in the diffractrogram. Nitrogen
adsorption experiments on the calcined Beta samples gave
an isotherm very similar to that of pure silica Beta with val-
ues of micropore volume of 0.20–0.21 cm3 g−1 and BET
surface areas of 450–475 m2 g−1. All Beta samples were
used directly after calcination or after activation for 2 h at
473 K under dynamic vacuum.
2.5. Synthesis of 4-(but-3-enoxy)-benzaldehyde (1k)
Following the procedure for aldehyde 1j, 4-hydroxyben-
zaldehyde (1.2 g, 10 mmol), 1.4 g of 4-bromo-1-butene,
30 mL of dry dimethylformamide,potassium iodide (0.080 g,
0.75 mmol), and 5.5 g (39 mmol) of fine powdered potas-
sium carbonate were employed. After evaporation of the
solvent 0.30 g (1.7 mmol, 17%) of the desired aldehyde 1j
was obtained.
2.2. In situ IR measurements
MS m/z (%): 176 (44) [M]+, 164 (39), 148 (31), 135
(13), 121 (54), 110 (100), 105 (13), 93 (20), 77 (17), 65 (20),
55 (86).
For the IR study of the interaction of the different aro-
matic aldehydes with active sites of zeolites, a sample of the
zeolite was first treated overnight at 673 K under dynamic
vacuum (1.33 × 10−3 Pa) in order to remove adsorbed H2O.
Then the corresponding aldehyde was adsorbed and the IR
spectrum was recorded at room temperature. Afterward, the
organic material was desorbed successively at 323, 373, and
473 K. An IR spectrum was recorded after each desorption.
2.6. Synthesis of 4-(but-3-enoxy)-benzene (8)
Following the procedure for aldehyde 1j, phenol (0.96 g,
10 mmol), 1.4 g (10 mmol) of 4-bromo-1-butene, 30 mL
of dry dimethylformamide, potassium iodide (0.17 g, 0.81
mmol), and 5.5 g (39 mmol) of fine powdered potassium car-
bonate were employed. After evaporation of the solvent the
crude reaction mixture was submitted to column chromatog-
raphy (4:1 mixture of hexane and tert-butyl methyl ether;
Rf = 0.8). 0.70 g (4.31 mmol, 41%) of the desired ether 8
were obtained.
2.3. Oxidations
Aldehdyes 1a to 1i, hydrogen peroxide (50%), meta
chloroperbenzoic acid (77% max), and solvents were pur-
chased from Aldrich in the highest purity available (ꢀ 98%)
and used without further purification. GC analyses were car-
ried out on a HP 5890 gas chromatograph equipped with a
25-m HP-5 column. GC-MS analyses for the identification
of products were carried out on an Agilent Technologies
6890N apparatus coupled with an Agilent Mass Selective
MS m/z (%): 162 (13) [M]+, 147 (10), 107 (12), 94
(100), 66 (19), 55 (14). 1H NMR (300 MHz, CDCl3, 298 K):
δ = 1.74 (s, 3 H, 8-H or 9-H), 1.79 (s, 3 H, 8-H or 9-H), 4.50
(d, 2 H, 5-H), 5.50 (t, J = 6.7 Hz, 1 H, 6-H), 6.94 (m, 3 H,
1-H, 3-H), 8.36 (t, 2 H, 2-H). 13C NMR (75 MHz, CDCl3,

A. Corma et al. / Journal of Catalysis 221 (2004) 67–76
69
298 K): δ = 18.6, 26.2, 65.0, 115.1, 120.2, 120.9, 129.8,
Ester 2g, MS m/z (%): 122 (31) [M]+, 94 (100), 77 (5),
136.4, 159.2.
66 (27), 51 (7).
Alcohol 3g, MS m/z (%): 94 (100) [M]+, 66 (30), 55 (5),
2.7. General procedure for the Baeyer–Villiger oxidation
50 (5).
Acid 4g, MS m/z (%): 122 (89) [M]+, 105 (100), 77 (62),
The amount of 0.50 g of aldehyde, 0.30 g of 50% aqueous
hydrogen peroxide, and water (if required) were dissolved
in 3.0 g of the corresponding solvent (normally acetonitrile).
Then 50 mg of the catalyst was added, and the reaction mix-
ture was stirred and heated to 353 K for 7 h. The reaction
was followed by gas chromatography, and the products were
identified by comparison with reference samples, by GC-MS
spectroscopy, or after purification by 1H NMR spectroscopy.
Reaction conditions for Tables 7 and 8: 0.25 g (1.4 mmol)
of aldehyde 1j or 1k, 0.15 g (2.3 mmol) of 50% aqueous
hydrogen peroxide, 1.5 g acetonitrile, 25 mg of catalyst, and
353 K reaction temperature.
51 (23).
2.8. Oxidation of aldehyde 1j with mCPBA
The amounts of 25 mg of aldehyde 1j and 9 mg of
mCPBA were dissolved in 150 mg of acetonitrile. After
standing at room-temperature for 2 h the sample was sub-
mitted to GC and GC-MS analyses.
Reaction conditions for Scheme 5 and Table 9: 0.25 g
(1.8 mmol) of aldehyde 1a, 0.16 g (2.3 mmol) 50% aqueous
H2O2 50%, 0.29 g (1.8 mmol) of ether 8, 1.5 g of acetoni-
trile, and 0.025 g of catalyst.
Ester 2j, MS m/z (%): 206 (2) [M]+, 138 (65), 110 (100),
81 (8), 69 (38), 53 (14), 41 (23).
Reaction conditions for Table 10: for the first entry, 1.0 g
of aldehyde 1a, 0.6 g of 50% aqueous hydrogen peroxide,
6.0 g of acetonitrile, 353 K, 7 h, and 100 mg of Sn-Beta; after
the reaction, the catalyst was filtered out, air-dried overnight,
and activated at 473 K/vacuum for 2 h as described for fresh
catalyst. For entries 2–5 and 7, the amount of reagents was
scaled down with respect to the catalyst amount available.
Ester 2a, MS m/z (%): 152 (37) [M]+, 124 (76), 109
(100), 81 (36), 53 (12).
Alcohol 3j, MS m/z (%): 178 (8), 110 (100), 81 (19), 69
(39), 53 (16).
Epoxyaldehyde 5j, MS m/z (%): 206 (67) [M]+, 147
(8), 136 (22), 121 (62), 105 (21), 85 (100), 71 (30), 59
(76), 51 (14), 43 (38). 1H NMR (300 MHz, CDCl3, 298 K):
δ = 1.38 (s, 3 H, 9-H or 10-H), 1.41 (s, 3 H, 9-H or 10-H),
ꢀ
3.17 (dd, J6,7 = 5.95 Hz, J6 ,7 = 5.95 Hz, 1 H, 7-H), 4.13
(dd, J6 ,7 = 5.95 Hz, J6 ,6 = 11.0 Hz, 1 H, 6ꢀ-H), 4.25 (dd,
ꢀ
ꢀ
ꢀ
J6,7 = 5.95 Hz, J6,6 = 11.0 Hz, 1 H, 6-H), 7.05 (d, J = 8.80
Alcohol 3a, MS m/z (%): 124 (92) [M]+, 109 (100), 81
(45), 53 (16).
Hz, 2 H, 4-H), 7.85 (d, J = 8.80 Hz, 2 H, 3-H), 9.89 (s, 1 H,
1-H).
Acid 4a, MS m/z (%): 152 (91) [M]+, 135 (100), 122
(9), 109 (9), 107 (11), 105 (11), 91 (22), 77 (18), 63 (11).
Ester 2b, MS m/z (%): 166 (24) [M]+, 138 (25), 110
(100), 81 (15), 53 (6).
Alcohol 3b, MS m/z (%): 138 (55) [M]+, 110 (100), 81
(15), 53 (6).
Ester 2c, MS m/z (%): 180 (49) [M]+, 152 (13), 110
(100), 81 (15), 65 (7), 53 (8).
Epoxyester 6j, MS m/z (%): 222 (48) [M]+, 194 (15),
138 (12), 123 (13), 110 (100), 93 (10), 85 (30), 71 (24), 65
(11), 59 (32), 43 (29).
Alcohol 3c, MS m/z (%): 152 (49) [M]+, 110 (100), 81
(15), 65 (8), 53 (8).
Epoxyphenol 7j, MS m/z (%): 194 (50) [M]+, 123 (8),
Ester 2d, MS m/z (%): 214 (48) [M]+, 186 (100), 157
(14), 129 (12), 109 (19), 77 (25), 51 (12).
1
110 (100), 85 (43), 59 (52), 43 (17). H NMR (300 MHz,
CDCl3, 298 K): δ = 1.35 (s, 3 H, 8-H or 9-H), 1.38 (s,
3 H, 8-H or 9-H), 3.17 (dd, signal overlapping partially
with signals of epoxyaldehyde 5j, 1 H, 6-H), 4.01 (dd,
Alcohol 3d, MS m/z (%): 186 (100) [M]+, 157 (12), 129
(8), 109 (12), 77 (10), 51 (5).
Alcohol 3e, MS m/z (%): 108 (100) [M]+, 90 (21), 80
(43), 77 (60), 63 (10), 51 (22).
J5 ,6 = 5.60 Hz, J5 ,5 = 10.9 Hz, 1 H, 5-H or 5ꢀ-H), 4.07
ꢀ
ꢀ
ꢀ
ꢀ
(dd, J5,6 = 5.60 Hz, J5,5 = 10.9 Hz, 1 H, 5-H or 5 -H), 6.76
(d, J = 9.20 Hz, 2 H, 2-H or 3-H), 6.82 (d, J = 9.20 Hz,
2 H, 2-H or 3-H).
Acid 4e, MS m/z (%): 136 (82) [M]+, 119 (69), 91 (100),
65 (19).
Ester 2f, MS m/z (%): 178 (17) [M]+, 163 (21), 135
(100), 107 (31), 91 (12), 77 (9), 55 (8).
Alcohol 3f, MS m/z (%): 150 (23) [M]+, 135 (100), 107
(31), 95 (11), 77 (8).
Acid 4f, MS m/z (%): 178 (19) [M]+, 164 (12), 163
(100), 135 (28), 115 (8), 91 (28), 77 (12).

70
A. Corma et al. / Journal of Catalysis 221 (2004) 67–76
Fig. 1. Yield of ester 2a and alcohol 3a versus total conversion of aldehyde
1a.
Scheme 1.
Table 1
2.9. Oxidation of aldehyde 1j with hydrogen peroxide
Influence of the water amount on the hydrolysis of the ester to the alcohol
catalyzed by methyltrioxorhenium (MTO) in acetonitrile
Entry Water TON Conversion
Product distribution
(mg)
(%)
Ester 2a Alcohol 3a Other
The amounts of 0.50 g of aldehyde 1j and 0.30 g of 50%
aqueous hydrogen peroxide were dissolved in 3.0 g of ace-
tonitrile. Then, 10 mg (1.56 mol%) of MTO was added, and
the reaction mixture stirred and heated to 353 K for 7 h and,
then analyzed by GC and GC-MS analyses.
1
2
3
4
5
6
7
0
20
50
100
200
300
500
317
314
295
310
350
307
327
56
57
53
56
62
54
59
54
52
39
32
18
15
4
46
48
61
68
82
85
96
0
0
0
0
0
0
0
2.10. Oxidation of aldehyde 1j with hydrogen peroxide
catalyzed by methyltrioxorhenium in methyl-tert-butyl ether
Reaction conditions: 0.5 g of aldehyde 1a, 0.30 g of 50% aqueous hydrogen
peroxide, 3.0 g of acetonitrile, water (when indicated), and 50 mg of catalyst
Beta-1 (Sn-Beta) were stirred for 7 h at 353 K.
To a solution of 0.195 g of aldehyde 1j in 1 mL of 3 M
etheric solution of hydrogen peroxide 1.3 mg MTO was
added. The reaction was stirred at room temperature for 7 h
and then analyzed by GC and GC-MS analyses.
total conversion is presented. It can be seen there that the
formate ester 2a is a primary and unstable product, whereas
the phenol 3a presents a secondary character. Indeed, prod-
uct 3a should be formed by hydrolysis of ester 2a with the
water present in the reaction media.
2.11. Oxidation of aldehyde 1k with mCPBA
The amounts of 25 mg of aldehyde 1k and 13 mg of
mCPBA were dissolved in 155 mg of acetonitrile. After
standing at room temperature for 2 h the sample was sub-
mitted to GC and GC-MS analyses.
It must be pointed out that in most of the BV oxidation
of aldehydes, the desired product is the alcohol rather than
the ester. However, since the alcohol is produced in a con-
secutive step by hydrolysis of the ester, it becomes clear that
a successful catalyst aimed to produce high yields of alco-
hol in a one-pot reaction should be able to perform the BV
reaction in the presence of water, while catalyzing rapidly
the hydrolysis of the ester formed in the first oxidation step.
Thus, in order to couple the two reactions, i.e., oxidation
and hydrolysis, the process was carried out introducing dif-
ferent amounts of water in the reaction media, and using first
the Sn-Beta catalyst. The results presented in Table 1 show
that the oxidation activity of the Sn-Beta zeolite is practi-
cally not affected by the presence of water. However, the rate
of hydrolysis of the ester 2a increases when increasing the
water content, reaching selectivities higher than 95% to the
methoxyphenol (3a). If the hydrolysis should be avoided the
reaction can be carried out with a deficit of hydrogen perox-
ide. Employing only 0.35 eq, the ester 2a is observed in a
relative selectivity of 90% (Table 2, entry 1). Furthermore,
under these conditions the hydrogen peroxide is consumed
almost completely. This indicates that competitive decom-
Ester 2k, MS m/z (%): 192 (15) [M]+, 164 (24), 136 (5),
81 (9), 55 (39).
Alcohol 3k, MS m/z (%), 164 (38) [M]+, 136 (7), 110
(100), 93 (6), 81 (11), 55 (26).
3. Results and discussion
In order to design the catalyst and reaction conditions
for the Baeyer–Villiger reaction of aldehydes, we selected
the para-anisaldehyde (1a) as model substrate. The starting
catalyst was a Sn-Beta sample (Beta-1) with 2 wt% of SnO2
content. The reaction was performed using H2O2 as oxi-
dant, acetonitrile as solvent, and 353 K reaction temperature.
After 7 h reaction time the para-anisaldehyde conversion
was 56% and the only observed products were the para-
methoxyphenyl formate (2a) and para-methoxyphenol (3a,
Scheme 1). In Fig. 1 the yield of reaction products versus

A. Corma et al. / Journal of Catalysis 221 (2004) 67–76
71
Table 2
BV oxidation of aldehyde 1a with different amounts of hydrogen peroxide (50%) in acetonitrile
Entry
H O
2
TON
Conversion 1a
Conversion H O
2 2
Product distribution
2
(eq.)
(%)
(%)
Ester 2a
Alcohol 3a
Other
1
2
3
4
5
0.35
0.45
0.80
1.0
169
199
242
285
317
31
36
44
51
56
88
80
57
51
46
90
89
83
72
54
10
11
17
28
46
0
0
0
0
0
1.2
Reaction conditions: 0.5 g of aldehyde 1a, 50% aqueous hydrogen peroxide (amount as indicated), 3.0 g of acetonitrile, and 50 mg of catalyst Beta-1 (Sn-Beta)
were stirred for 7 h at 353 K.
Table 3
Influence of the solvent on the hydrolysis of the ester to the alcohol
Entry
Water
(mg)
Solvent
TON
Conversion
(%)
Product distribution
Ester 2a
Alcohol 3a
other
1
2
3
4
5
6
7
8
–
500
–
500
–
500
–
500
Acetonitrile
Acetonitrile
Dioxane
Dioxane
Ethanol
Ethanol
DMSO
DMSO
317
327
259
106
317
244
5
56
59
46
19
57
44
1
54
4
77
54
1
5
46
55
46
96
23
46
99
95
54
45
0
0
0
0
0
0
0
0
5
1
Reaction conditions: 0.5 g of aldehyde 1a, 0.30 g of 50% aqueous hydrogen peroxide, 3.0 g of solvent, water (when indicated), and 50 mg of catalyst Beta-1
(Sn-Beta) were stirred for 7 h at 353 K.
position of hydrogen peroxide does not occur and the selec-
tivity with respect to the oxidant is excellent.
the best conditions the ester 2a can be obtained with a 77%
selectivity at a conversion of 46% (Table 3, entry 3).
When DMSO was used as solvent, the activity of the cata-
lysts was very low, achieving less than 5% of conversion
after 7 h of reaction time (Table 3, entry 7). This result might
be attributed to the strong adsorption of the DMSO on the
active Sn centers through the oxygen of the sulfoxide group.
The effect of the water concentration was studied with
the different solvents and it was found that in the case of the
protic solvent ethanol, the water addition decreases the con-
version significantly and has a minor effect on the rate of
hydrolysis (Table 3, entries 5 and 6), while using ethanol as
solvent the hydrolysis was already completed without fur-
ther water addition. In the case of the polar aprotic solvent
dioxane, the catalyst activity is also lowered when increasing
water concentration (Table 3, entries 3 and 4). For the prod-
uct selectivity an increase of hydrolysis was observed and at
longer reaction times the product ratio was shifted more and
more toward phenol 3a as it has been described already for
acetonitrile.
3.1. Influence of the solvent for the Sn-Beta catalyst
If one takes into account that in the present reaction two
molecules with different polarities, i.e., H2O2 (aq) and para-
anisaldehyde have to diffuse through the zeolite pores and
adsorb on the catalyst surface, it becomes clear that de-
spite the catalyst properties the polarity of the solvent should
also play an important role on the final catalytic behavior.
In order to study this, the Baeyer–Villiger rearrangement of
para-anisaldehyde (1a) was carried out in four different sol-
vents: acetonitrile, dioxane, ethanol, and dimethyl sulfoxide
(DMSO). The results are summarized in Table 3 and they
show that when one goes from the less polar solvent, diox-
ane, to the more polar acetonitrile and ethanol, the activity of
the catalyst increases. Furthermore, the selectivity to prod-
ucts 2a and 3a is strongly affected by the nature of the
solvent, in such a way that the phenol selectivity increases
in the same order as the polarity of the solvent: ethanol >
acetonitrile > dioxane. In fact, when using ethanol as sol-
vent, it is possible to obtain phenol 3a with 99% selectivity
(Table 3, entry 5). These findings indicate that the transition-
state complex (TSC) which leads to the ester as well as the
one which leads to the ester hydrolysis is better stabilized
when the solvent polarity is higher, decreasing the free en-
thalpy of the TSC and enhancing the reaction rate of both
processes. On the other hand, when the ester 2a is the de-
sired product dioxane should be the solvent of choice. Under
From these results we can conclude that the selectivity to
ester 2a and phenol 3a can be modified by using the adequate
solvent. In dioxane the ester 2a is the predominant product
whereas in ethanol or aqueous acetonitrile the phenol 3a is
obtained almost exclusively.
3.2. Nature and optimization of the catalytic active sites
We have seen above the influence of solvent and H2O
content on the relative selectivity to ester and alcohol. How-

72
A. Corma et al. / Journal of Catalysis 221 (2004) 67–76
Table 4
Different Al and Sn contents in Beta catalysts in the Baeyer–Villiger oxidation of para-anisaldehyde (1a) with hydrogen peroxide (50%) in acetonitrile
Entry
Catalyst
Conversion
(%)
Product distribution
SnO (wt%)
Si/Al ratio (mol/mol)
2a
3a
Other
0
2
1
2
3
4
5
6
7
8
Beta-1
Beta-2
Beta-3
Beta-4
Beta-5
Beta-6
Beta-7
Beta-8
2.0
2.6
2.4
0.0
0.0
0.0
0.0
0.0
> 10,000
100
28
> 10,000
250
56
74
78
3
54
11
5
84
17
8 (51)
1 (47)
11
46
87
89
16
a
2
6
a
0
0
84
83
a
100
30
15
79 (61)
87 (70)
83
79 (39)
95 (50)
86
12 (10)
a
4 (3)
3
Reaction conditions: 0.5 g of aldehyde 1a, 0.30 g of 50% aqueous hydrogen peroxide, 3.0 g of acetonitrile, and 50 mg of a catalyst sample were stirred for 7 h
at 353 K. Values in parentheses were obtained after 1 h reaction time.
a
Mainly anisic acid 4a.
ever, there is another way to modify the relative rate of the
oxidation and hydrolysis reaction, namely by introducing
Brønsted acid sites into the Lewis-type Sn-Beta zeolite cata-
lyst. Indeed, the presence of Brønsted acidity in the catalyst
should increase the rate of hydrolysis and therefore should
increase the yield of alcohol. In order to study this hy-
pothesis a series of zeolites with very close amounts of Sn
but with different framework aluminum content were syn-
thesized. If one takes into account that an acidic bridging
hydroxyl group should be formed per each framework alu-
minum, we have prepared catalysts with different ratios of
oxidating (Sn) to hydrolysis (H+) active sites (see Table 4).
When these catalysts were tested for the BV reaction of
para-anisaldehyde, it could be seen that the yield of the alco-
hol (3a) increases when increasing the framework aluminum
Fig. 2. Adsorption and desorption of para-anisbenzaldehyde measured by
in the Sn-Beta zeolite. What appears more interesting is that
conversion is enhanced when increasing the Brønsted acid-
ity of the zeolite, indicating that the bridging hydroxyls can
also be active sites for the BV oxidation of aldehydes. Thus,
we have prepared a series of Sn-free Al-Beta zeolites with
different framework Si/Al ratios, and they show an excellent
catalytic activity. However, it appears that Al-Betas are glob-
ally less selective than Sn-Beta since the former produces
larger amounts of subproducts, specially anisic acid, indi-
cating that they also catalyze the migration of the H atom
besides the migration of the aromatic moiety. Note that this
lower selectivity is not an effect of the conversion level, since
when similar levels of conversion are compared (Table 4, en-
tries 1, 6, and 7), the amounts of subproducts produced with
Sn-Beta are clearly lower.
“in situ” IR spectroscopy. In the case of Sn-Beta (Beta-1) the displayed
spectrum was obtained after desorption at 373 K and in the case of Si-Beta
(Beta-4) after desorption at 323 K.
case, Sn-Beta does not activate the H2O2 but the carbonyl
group by coordination with the Lewis-acidic Sn center.
In order to see if the same reaction model applies to
the oxidation of aldehydes, we have adsorbed a comparable
amount of para-anisaldehyde on pure silica Beta (Beta-4)
and on Sn-Beta (Beta-1) samples in a vacuum IR cell. It can
be seen in Fig. 2 that for Si-Beta, the carbonyl band appears
at 1696 cm−1 which corresponds to the physically adsorbed
para-anisaldehyde. When the sample was heated at 373 K
the product desorbs and the 1696 cm−1 band disappears. In
the case of the Sn-Beta catalyst, after heating at 373 K a band
at 1588 cm−1 remains. This corresponds to adsorbed alde-
hyde in which the 1696 cm−1 band is shifted to 1588 cm−1
owing to the interaction of the carbonyl band of the alde-
hyde with the Sn Lewis acid center. This result indicates that
also in the case of aromatic aldehydes an activation of the
carbonyl group by the Sn sites occurs that facilitates the nu-
cleophilic attack of the H2O2 to the corresponding carbon
atom.
3.3. Mechanism of the reaction and the effect of the
substituent in the aromatic ring
In the study of the Baeyer–Villiger oxidation of cyclic ke-
tones to lactones using the system Sn-beta/H2O2 we have
shown, by means of 18O-labeled 2-methylcyclohexanone,
that the mechanism of the Baeyer–Villiger oxidation with
hydrogen peroxide proceeds through a Criegee adduct of the
hydrogen peroxide with the ketone substrate [15,16]. In that
If the above is true, the electron donor properties of differ-
ent substituents in the benzene ring should have an influence
on conversion and selectivity. From the results presented

A. Corma et al. / Journal of Catalysis 221 (2004) 67–76
73
Table 5
Baeyer–Villiger oxidation of benzaldehydes involving different para sub-
stituents with hydrogen peroxide catalyzed by Sn-Beta
Entry
Substrate
Conversion
(%)
Product distribution
2 + 3 Other
R
4
1
2
3
4
5
6
7
1a
1b
1c
1d
1e
1f
OMe
OEt
OnPr
OPh
Me
56
69
82
70
32
13
19
> 99
> 99
> 99
> 99
72
0
0
0
0
0
0
5
0
0
0
0
28
tBu
H
52
43
1g
0
> 99
Reaction conditions: 0.5 g of aldehyde 1, 0.30 g of 50% aqueous hydrogen
peroxide, 3.0 g of acetonitrile, and 50 mg of catalyst Beta-1 (Sn-Beta) were
stirred for 7 h at 353 K.
Fig. 3. IR spectra for Beta-1 (spectra a, c, and e) and Beta-4 (spectra b, d,
and f) samples after adsorption and desorption of para-methylbenzaldehyde
(1e, spectra a and b) and of benzaldehyde (1g, spectra c to f). The desorption
temperature was 323 K for spectra a, b, e, and f, and 298 K for spectra c
and d.
zophenones with peracids [20]. However, it has also been
pointed out that not only the migration ability is modified by
different aromatic substituents but also the electronic nature
of the carbonyl bond and therewith the nucleophilic attack of
the hydrogen peroxide onto the carbonyl carbon atom [20].
Indeed, from in situ IR measurements this effect might be
confirmed here. Fig. 3 shows the corresponding IR spectra
of methylbenzaldehyde 1e and benzaldehyde 1g. The car-
bonyl band for both compounds (physisorbed) can be found
at 1705 versus 1696 cm−1 observed for physisorbed para-
methoxybenzaldehyde (cf. Fig. 3). The difference is even
more pronounced for the chemisorbed derivatives. Here, the
value of ca. 1635 cm−1 for aldehydes 1e and 1g (Fig. 3)
is changed to 1588 cm−1 for aldehyde 1a (Fig. 2). In sum-
mary, it can be concluded that electron donating substituents
in the para position of the benzaldehyde enhance the reac-
tion rate by favoring the nucleophilic attack on the carbonyl
group and lowering the activation energy for the migration
of the aromatic moiety. Both effects are beneficial for the
selectivity toward the ester and alcohol (cf. Scheme 2, left
pathway) and, consequently, the ester is the only primary
product observed in the case of alkoxy benzaldehydes with
a pronounced positive mesomeric effect.
Scheme 2. Baeyer–Villiger reaction of aromatic aldehydes 1. The aldehyde
1 and the oxidant first form the “Criegee Intermediate,” and then migration
of the aromatic ring leads to the formate ester 2 which can be hydrolyzed to
the phenol 3 (left-hand side). Migration of the hydrogen atom in the Criegee
Intermediate yields the corresponding carboxylic acids 4 (right-hand side).
in Table 5 it becomes apparent that conversion decreases
when the electron donor capacity of the group in para po-
sition decreases. A parallel effect is observed for the product
selectivity which is shifted more and more from 100% mi-
gration of the aromatic moiety for the alkoxy sustituents (cf.
Scheme 2, left-hand side) to yield ester 2 and alcohol 3, to
proton migration which yields the acid 4 (Scheme 2, right-
hand side). It seems that the migration ability of the aromatic
unit is improved by electron-donating substituents in para
position enhancing the reaction rate and favoring its migra-
tion over the alternative hydrogen migration. A similar effect
has been found for the BV reaction of mono-substituted ben-
Finally, the influence of the position of a methoxy group
in the aromatic ring on the catalyst activity was studied.
The highest catalytic activity is found for aldehyde 1a with
the substituent in para position (Table 6, entry 1), followed
by the ortho isomer (Table 6, entry 3). However, when the
methoxy group is placed in the meta position, a low conver-
sion is achieved, being meta-methoxy benzoic acid (4h) the
main product detected in the reaction mixture. This result is
not surprising having in mind that the mesomeric effect of
the meta substituent cannot influence the electronic nature
of the carbonyl bond or of the migrating atom. The lower
conversion for the ortho aldehyde versus the para isomer is

74
A. Corma et al. / Journal of Catalysis 221 (2004) 67–76
Table 6
Baeyer–Villiger oxidation of para, meta, and ortho-methoxybenzaldehyde with hydrogen peroxide catalyzed by Sn-Beta
Entry
Aldehyde
TON
Conversion
(%)
Product distribution
Ester 2 phenol 3
Acid 4
Other
1
2
3
para
meta
ortho
1a
1h
1i
291
70
134
56
12
46
> 99
0
> 99
0
81
0
0
19
0
Table 7
Chemoselectivities for different oxidation methods in the Baeyer–Villiger
oxidation of para-(3-methylbut-2-enoxy)-benzaldehyde (1j)
Entry Oxidation system Conversion
(%)
Product distribution
2j + 3j 5j 6j + 7j Other
a
b
1
2
3
4
5
H O , Beta-1
61
0
15
66
20
85
0
0
0
0
15
2
2
H O , Beta-7
0
0
0
0
21
32
31
2
2
c
H O , MTO
75
62
12
4
6
29
2
2
d
H O , MTO
2
2
e
mCPBA
28
a
b
1:4 ratio of ester 2j to alcohol 3j.
Unidentified higher oxida-
c
d
Scheme 3.
tion products excluding epoxide 5j.
In acetonitrile.
Reaction con-
ditions similar as in Ref. [23]; a etheric solution of hydrogen peroxide is
e
used.
Only 0.3 eq of the oxidant was employed.
Table 8
Chemoselectivities for different oxidation methods in the Baeyer–Villiger
oxidation of para-(3-butenoxy)-benzaldehyde (1k)
Entry Oxidation system Conversion
(%)
Product distribution
2k + 3k 5k 6k + 7k Other
a
1
2
3
H O , Beta-1
41
0
22
> 99
0
> 99
0
0
0
0
0
0
0
0
0
2
2
H O , Beta-7
2
2
b
mCPBA
a
b
Scheme 4.
2:3 ratio of ester 2k and alcohol 3k.
Only 0.4 eq of oxidant was
employed.
probably due to steric hindrance of the two neighbored sub-
stituents which might render more difficult the formation of
the transition state. A similar trend has been observed for the
BV oxidation with peracids [20].
systems and they are compared with those obtained when
methyltrioxorhenium/H2O2 and meta-chloroperbenzoicacid
are used as oxidants. It can be seen that mCPBA gives a
mixture of all possible oxidation products from BV oxida-
tion and epoxidation (Table 7, entry 5 and Scheme 3). The
MTO/H2O2 oxidant shows a high epoxidation activity but
less BV activity. Consequently, the BV products 2j and 3j
are not observed and, if formed, they are oxidized further to
the corresponding epoxidized BV products 6j and 7j which
are the by-products for the epoxide 5j (Table 7, entries 3
and 4). Sn-Beta (catalyst Beta-1) is active and selective for
the BV oxidation yielding mainly compounds 2j and 3j (Ta-
ble 7, entry 1), whereas, surprisingly, Al-Beta (Beta-7) was
not active at all (Table 7, entry 2). When aldehyde 1k was
used as substrate (Table 8 and Scheme 4) we found sim-
ilar results using both zeolites as catalysts. Thus, whereas
Sn-Beta yielded exclusively the ester 2k and the phenol 3k
(Table 8, entry 1), the Al-Beta was an inactive catalyst (Ta-
ble 8, entry 2). The peracid yields exclusively compounds
2k and 3k in this case probably due to the low activation
3.4. Synthesis of phenols with industrial interest
It has been reported that the phenols 4-(3-methyl-2-
butenoxy)-phenol(3j, Scheme 3) and 4-(3-butenoxy)-phenol
(3k, Scheme 4) are intermediates for the production of a
thermotropic liquid crystal polymers used in cosmetic com-
positions for human skin and/or lips [21], and for the synthe-
sis of effective insecticides against houseflies [22], respec-
tively. They might be produced by the BV oxidation of 4-(3-
methyl-2-butenoxy)-benzaldehyde (1j) and 4-(3-butenoxy)-
benzaldehyde (1k). However, both starting molecules have
in common a para substituent bearing a double bond. Thus,
the BV oxidation of those molecules will require not only
active but also chemoselective oxidants since the epoxida-
tion of the double bond must be avoided.
In Table 7 the results for the BV oxidation of aldehyde 1j
are presented using the Sn-Beta and Al-Beta/H2O2 oxidation

A. Corma et al. / Journal of Catalysis 221 (2004) 67–76
75
Table 10
Recycling of the Sn-Beta catalyst without treatment in the Baeyer–Villiger
oxidation of para-anisaldehyde (1a) with hydrogen peroxide (50%)
Entry
Special treatment
Cycle
TON
Conversion (%)
1
2
3
4
5
1
2
3
4
282
199
237
222
282
51
36
43
38
51
Recalcination
Extraction
4/1
6
7
1
1/1
277
207
50
37
Scheme 5.
The selectivity toward ester 2a and alcohol 3a was always > 98%.
3.5. Deactivation and regeneration of Sn-Beta
Table 9
Oxidation of para-anisaldehyde in the presence and absence of the potential
poison molecule 8
In order to study the catalyst deactivation during the BV
oxidation, we have carried out four reaction cycles with
para-anisaldehyde using Beta-1 as catalyst and acetonitrile
as solvent. At the end of each reaction the catalyst sample
was filtered out and submitted to the next run after activa-
tion at 473 K in vacuum. In Table 10 the results for Sn-Beta
(Beta-1) obtained after four cycles are presented showing
a decrease in activity from the first to the second cycle
(Table 10, entry 2), with no further decrease in the subse-
quent cycles. The loss of activity might be mainly due to the
block of the active sites and/or channels by organic mate-
rial. In fact, when, the used catalyst was soxhlet-extracted
with ethanol or acetone as solvent, aldehyde 1a and ester 2a
could be detected in the extract. However, when the soxhlet-
extracted catalyst was reused, the activity was not improved
(cf. Table 10, entries 2 and 7). The TG analysis of this ex-
tracted zeolite sample showed that 4 wt% of organic material
still remained on the catalyst.
Entry
Catalyst
Alkene 8
(eq)
Conversion
(%)
2a
3a
a
b
1
2
3
4
Beta-1
0
1.0
0
56
19
87
2
54
13
1
46
87
95
21
Beta-1
Beta-7
c
Beta-7
1.0
79
a
b
c
With respect to anisaldehyde.
From Table 1.
From Table 4.
of the monosubstituted double bond toward the epoxidation
(Table 8, entry 3).
The lack of activity of Al-Beta zeolite could be attributed
to a preferential interaction of the substrate with the Brøn-
sted acid sites through the double bond instead of through
the carbonyl group. Then, deactivation may derive from two
different effects:
IR spectroscopy on the catalyst after the reaction without
any treatment confirmed the presence of physisorbed alde-
hyde 1a. A temperature treatment of 473 K in vacuum was
able to remove the aldehyde completely and a spectrum of
a single compound (sharp bands) was observed (spectrum
not shown). Three characteristic bands can be attributed very
clearly the to ester 2a, namely 1727 cm−1 for the carbonyl
group of a formate ester, and 1602 and 1508 cm−1 for the
aromatic moiety. All these data indicate that there is still
some ester 2a adsorbed in a used catalyst sample which was
not removed by activation at 473 K. However, when the cata-
lyst, used in four cycles, was regenerated by calcination at
853 K for 3 h in air, the initial activity was fully recovered
(cf. Table 10, entries 1 and 5). When the same reaction pro-
tocol was carried out under reaction conditions that favor
the hydrolysis of the ester and the production of alcohol, the
catalytic performance of the catalyst was very similar.
(i) formal replacement of catalytically active protons by
fairly stable unreactive carbocations and therewith neu-
tralizing the active sites for the BV oxidation, and
(ii) blocking of the pores by these strongly adsorbed bulky
reaction intermediates [24,25].
In order to check the hypothesis presented above, we
did an additional experiment where the BV oxidation of
the para-anisaldehyde was carried out in the presence of 3-
methyl-2-butenoxybenzene (8) using Beta-1 and Beta-7 as
catalysts (Scheme 5). It can be seen that for Sn-Beta the
conversion decreases in the presence of this potential poison
molecule (Table 9, entries 1 and 2), probably due to compet-
itive adsorption and diffusion within the catalyst channels.
However, Beta-7 suffers by far stronger deactivation with
only 2% of conversion of aldehyde 1a after 7 h reaction
time (Table 9, entries 3 and 4). These results allow us to
conclude that deactivation observed for the Al-Beta zeolite
is due to the interaction of the Brønsted acid sites with the
double bond of the substrate. In contrast for the Lewis-acidic
Sn sites, a significant deactivation in the presence of double
bond has not been detected.
4. Conclusions
The heterogeneous oxidation systems Sn-Beta/H2O2 and
Al-Beta/H2O2 are both active for the Baeyer–Villiger oxida-

76
A. Corma et al. / Journal of Catalysis 221 (2004) 67–76
tion of aromatic aldehydes. With alkoxy substituents in ortho
and especially in para position the corresponding formate
ester is the primary reaction product with excellent selectiv-
ity. The highest selectivity toward the ester is obtained with
the Sn-Beta catalyst in dioxane as solvent. High selectivities
of the corresponding phenol can be obtained by employing
ethanol or aqueous acetonitrile as solvent with Sn-Beta. Al-
Beta is more efficient as catalyst for both Baeyer–Villiger
oxidation and ester hydrolysis and, therefore, achieves high
yields for the corresponding alcohol. However, this is a less
selective catalyst than Sn-Beta zeolite. It must be pointed
out that Sn-Beta is a chemoselective catalyst when the alde-
hyde reactant contains olefinic groups in an alkyl chain. With
these reactants the Al-zeolite gives no activity. Other clas-
sical BV oxidants such as peracids also yield the epoxida-
tion products. Thus Sn-Beta catalysts open a new entry to
aromatic phenols with unsaturated alkyl substituents which
are interesting intermediates in the chemical industry. The
mechanism for the Baeyer–Villiger oxidation of aromatic
aldehydes catalyzed by Sn-Beta is supposed to be similar
to the one for cyclic ketones and start with the activation of
the carbonyl bond by coordination to the Lewis acidic center
followed by a nucleophilic attack of the hydrogen peroxide
onto the more electrophilic carbonyl carbon atom.
References
[1] T. Saito, T. Hirayama, S. Sakagushi, JP 08151343, 1996.
[2] T. Sato, H. Matsumoto, T. Kakegawa, Y. Niino, JP 05301836, 1993.
[3] Y. Nihro, H. Furukawa, S. Sogawa, T.C. Wang, H. Miyataka, H. Mat-
sumoto, T. Miki, T. Satoh, Chem. Pharm. Bull. 42 (1994) 576–579.
[4] K. Schwabe, J. Redslob, D. Breitfeld, R. Zeisig, B. Tschiersch, W.
Wohlrab, K.D. Wozniak, C. Bayer, C. Nowak et al., DD 287482, 1991.
[5] P.A. Riley, C.J. Cooksey, C.I. Johnson, E.J. Land, A.M. Latter, C.A.
Ramsden, Eur. J. Cancer 33 (1997) 135–143.
[6] U. Wellnitz, S. Potsch, C. Garbe, G. Lassmann, Melanoma Res. 7
(1997) 288–298.
[7] G. Laßmann, S. Pötsch, DE 4344645, 1995.
[8] M.T. Caproiu, A.A. Banciu, E. Olteanu, RO 105090, 1994.
[9] G.R. Krow, Org. React. 43 (1993) 251–798.
[10] I.M. Godfrey, M.V. Sargent, J.A. Elix, J. Chem. Soc. Perkin Trans. 1
(1974) 1353–1354.
[11] N. Anoune, H. Hannachi, P. Lantéri, R. Longeray, C. Arnaud, J. Chem.
Ed. 75 (1998) 1290–1293.
[12] L. Syper, Synthesis (1989) 167–172.
[13] M. Matsumoto, H. Kobayashi, Y. Hotta, J. Org. Chem. 49 (1984)
4740–4741.
[14] R.H. Dodd, M. Le Hyaric, Synthesis (1993) 295–297.
[15] A. Corma, L.T. Nemeth, M. Renz, S. Valencia, Nature 412 (2001) 423–
425.
[16] M. Renz, T. Blasco, A. Corma, V. Fornés, R. Jensen, L. Nemeth,
Chem. Eur. J. 8 (2002) 4708–4717.
[17] A. Corma, M.T. Navarro, L. Nemeth, M. Renz, Chem. Commun.
(2001) 2190–2191.
[18] S. Valencia, A. Corma, UOP LCC, US 5968473A, 1999.
[19] A. Mann, C. Muller, E. Tyrrell, J. Chem. Soc. Perkin Trans. 1 (1998)
1427–1438.
[20] M. Renz, B. Meunier, Eur. J. Org. Chem. (1999) 737–750.
[21] F. Tournilhac, WO patent 0238112, 2001.
Acknowledgments
[22] C.A. Henrick, EP patent 169169, 1985.
The authors thank the CICYT (MAT2000-1392) and the
Generalitat Valenciana (CTIDIB/2002/16) for financial sup-
port. M.R. is grateful to the Spanish Ministry of Science and
Technology for a Ramón y Cajal fellowship.
[23] A.M.F. Phillips, C. Romão, Eur. J. Org. Chem. (1999) 1767–1770.
[24] A. Corma, H. García, J. Chem. Soc. Dalton Trans. 9 (2000) 1381–
1394.
[25] J.F. Haw, Phys. Chem. Chem. Phys. 4 (2002) 5431–5441.