Immobilised porphyrins in monoterpene photooxidations

Article abstract of DOI:10.1016/j.jcat.2008.04.001

Porphyrins were covalently linked to modified Merrifield polymers by chlorosulphonation activation of the porphyrin nucleus. These supported porphyrins were used as photosensitizers to promote singlet oxygen oxidation of monoterpenes with an efficiency that depends on porphyrin structure and the spacer used to link it to the polymer structure. The performance of these photosensitizers was studied. Citronellol and α-terpinene gave the expected singlet oxygen ene addition products. α-Pinene and β-pinene also gave products from non-ene reactions, which is explained by the existence of an alternative radical pathway.

Full text of DOI:10.1016/j.jcat.2008.04.001

Journal of Catalysis 256 (2008) 331–337
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Immobilised porphyrins in monoterpene photooxidations
∗
Sonia M. Ribeiro, Arménio C. Serra, A.M.d’A. Rocha Gonsalves
Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade de Coimbra, Rua Larga, P-3000 Coimbra, Portugal
a r t i c l e i n f o
a b s t r a c t
Article history:
Porphyrins were covalently linked to modified Merrifield polymers by chlorosulphonation activation of
the porphyrin nucleus. These supported porphyrins were used as photosensitizers to promote singlet oxy-
gen oxidation of monoterpenes with an efficiency that depends on porphyrin structure and the spacer
used to link it to the polymer structure. The performance of these photosensitizers was studied. Cit-
ronellol and α-terpinene gave the expected singlet oxygen ene addition products. α-Pinene and β-pinene
also gave products from non-ene reactions, which is explained by the existence of an alternative radical
pathway.
Received 12 February 2008
Revised 24 March 2008
Accepted 3 April 2008
Available online 9 May 2008
Keywords:
Polymer-supported porphyrins
Singlet oxygen
© 2008 Elsevier Inc. All rights reserved.
Monoterpene oxidation
1. Introduction
2. Experimental
Singlet oxygen generated from photoexcitation of ground-state
triplet oxygen is a convenient reagent for preparing endoperox-
ides or allylic alcohols [1–3]. Limitations of these reactions using
a sensitizer catalyst include the difficulty in isolating the product
from the catalyst and the poor stability of the catalyst for large-
scale reactions. The incorporation of the sensitizer in an insoluble
matrix is a convenient approach to overcome these problems. If
2.1. General
All solvents were purified before use according to procedures
described in the literature. α-Terpinene (85% purity), citronel-
lol (95% purity), α-pinene (98% purity), β-pinene (99% purity),
2,6-di-tert-butyl-4-methylphenol (BHT), Merrifield polymer 1%
cross-linked 200–400 mesh, 1,12-diaminododecane, and triphenyl-
posphine were used as purchased from Aldrich. Silicagel type 60
with particle size of 0.035–0.070 μm was purchased from Acros
Organics. Porphyrins 2, 3, 4, and 6 were prepared by condensation
of pyrrol and the corresponding aldehydes by the nitrobenzene
method [13]. Dodecylamino-modified Merrifield polymer (DAP)
was prepared as described previously [12]. 1H NMR spectra were
recorded on a 300-MHz Bruker-AMX spectrometer. All J values are
given in Hz. Mass spectra were obtained on a HP 5973 MSD appa-
ratus by electron impact at 70 eV. Elemental analyses were carried
out on a Fisons Instruments EA1108-CHNS-0 apparatus. Absorp-
tion spectra were measured on a Hitachi U-2001 spectrometer.
Gas chromatography was carried out using a Supelcowax (30 m ×
1
the ability to generate O2 is not overly affected, then the catalyst
is simultaneously protected from destruction, and the efficiency is
usually increased [4]. Due to their ability to generate singlet oxy-
gen and their resistance to degradation, porphyrins are known as
good photosensitizers in the case of homogeneous reactions [5],
but the linkage of porphyrins to polymeric matrices to take ad-
vantage of the heterogenation has not been fully and adequately
exploited to date. Some work on the incorporation of porphyrins
in polymeric matrices to be used as photosensitizers has been
described, including the use of a PEG-supported hydroxyphenyl
porphyrin [6], a copolymer of styrene and divinylbenzene doped
with porphyrins [7] or with adsorbed porphyrin [8], a cationic
functionalized polystyrene with ionic porphyrins [9], a hydrogel-
bound hematoporphyrin [10], and polystyrene microchannel chips
with silica-supported porphyrin [11].
We recently reported efficient singlet oxygen oxidations using
heterogeneous photosensitizers prepared by the covalent immobil-
isation of porphyrins on Merrifield-modified polymers [12]. In this
work, we extend our initial study, comparing the activity of new
immobilised porphyrins and testing their photooxidation efficiency
on monoterpenes.
0.25 mm) capillary column on a Hewlett–Packard 5890A instru-
ment with a Hewlett–Packard 3396A integrator. Gas chromatogra-
◦
◦
−1
◦
phy (GC) analyses were run at 50 C (5 min)/10 C min /200 C
20 min) for the α-pinene and β-pinene oxidation experiments
(
◦
◦
−1
◦
and at 80 C (2 min)/20 C min /200 C (20 min) for the citronel-
◦
lol oxidation experiment, at a detector temperature of 250 C and
an injector temperature of 220 C.
◦
2
.2. Preparation of the photosensitizers (PS1–PS5)
The PS1 photosensitizer was prepared as described previ-
Corresponding author. Fax: +351 239 826069.
E-mail address: arg@qui.uc.pt (A.M.d’A. Rocha Gonsalves).
ously [12]. The polymeric photosensitizers PS1 to PS4 were pre-
pared by the following general procedure. At room temperature,
*
0
021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.04.001

332
S.M. Ribeiro et al. / Journal of Catalysis 256 (2008) 331–337
1
5 mL of chlorosulphonic acid was added to 213 mg of each por-
dimethyloct-5-en-1-ol (10) and 6-hydroperoxy-3,7-dimethyloct-7-
en-1-ol (11), was estimated by 1H NMR [16].
phyrin (2 to 4). The solution was stirred for 2 h and then carefully
poured over ice to precipitate the porphyrin. The precipitate was
filtered, dried, and dissolved with dichloromethane, and the re-
sulting solution was dried with sodium sulphate. The solution was
concentrated to 30 mL, after which 10 mL of pyridine was added,
followed by 376 mg of the aminoalkylated polymer (DAP). The
◦
mixture was stirred overnight at 30 C, filtered, and washed with
dichloromethane, tetrahydrofuran, methanol, and dichloromethane
again. Nonbonded porphyrin was eliminated with these washings.
After the solid was dried under vacuum, elemental analysis was
carried out to determine the porphyrin incorporation in each of
the polymeric photosensitizers PS2 to PS4.
(
10)
(11)
2.3.6. (E)-7-Hydroperoxy-3,7-dimethyloct-5-en-1-ol (10)
1
For PS5, the same procedure was followed but using the same
amount of Merrifield polymer instead of DAP.
H NMR (300 MHz, CDCl ): δ = 0.87 (d, J = 6.4 Hz, 3H,
3
H
CHCH3), 1.25 ± 1.71 (m, 5H, CH2CHCH2), 3.64 (m, 2H, CH2OH);
1
5
.29 [s, 6H, COOH(CH3)2], 5.55 (d, J = 15.8 Hz, 1H, COOHCH=CH),
2
2
.3. Photooxidation experiments
.63 (m, 1H, COOHCH=CH).
.3.1. General procedure
2
.3.7. 6-Hydroperoxy-3,7-dimethyloct-7-en-1-ol (11)
Photooxidation experiments were carried out at room tempera-
δH = 0.85 (d, J = 6.5 Hz, 3H, CHCH3), 1.25 ± 1.71 (m, 5H,
ture using a laboratory-built photoreactor consisting of three halo-
gen 50 W lamps regularly placed around the reaction flask. The
reactions were done in a 100-mL flask equipped with a water con-
denser and an air inlet. The solutions in CHCl3 were irradiated,
with a stream of air continuously flowing through the flask. Then
the reaction mixture was filtered to recover the sensitizer, and
the filtrate was evaporated to dryness. Typical experiments are de-
scribed for different substrates with a sensitizer-to-substrate ratio
of 1/5000.
CH2CHCH2), 3.64 (m, 2H, CH2OH), 1.65 (s, 3H, CqCH3), 1.95 (m,
2
H, COOHCH2), 4.24 (m, 1H, CHOOH), 4.95 (s, 2H, Cq=CH2).
The proportion of the two regioisomers (10) and (11) was esti-
mated by 1H NMR: % for (10) = [(areaδ =5.55 + areaδ =5.68)/2] ×
H
H
1
00/[(areaδ =5.55 + areaδ =5.68)/2 + (areaδ =4.95)/2].
H
H
H
2.3.8. α-Pinene (12) and β-pinene (13)
The substrate (9.8 mmol) in 130 mL of chloroform was mixed
−3
with the appropriate amount of photosensitizer (1.7 × 10
mmol
2.3.2. α-Terpinene (7)
of porphyrin or supported porphyrin) to originate a sensitizer-
to-substrate molar ratio of 1/5000, after which 406 mg of base
(sodium hydrogen carbonate) was added. The evolution of the re-
action was monitored by GC. In the end, the reaction mixture was
submitted to reduction with triphenylphosphine, and the products
were isolated, as a single fraction, by column GC on silica using
CH Cl as eluent. Analysis of the product by 1H NMR spectroscopy
The substrate (4.9 mmol) in 65 mL of chloroform was mixed
−
4
with the appropriate amount of photosensitizer (9.8 × 10
mmol
of porphyrin or supported porphyrin) to originate the 1/5000 mo-
lar ratio of sensitizer to substrate, and 203 mg of base (sodium
hydrogen carbonate) was added. The evolution of the reaction was
monitored by UV–vis spectroscopy at 268 nm, with disappearance
of the reagent verified by GC. The photosensitizer was collected
by filtration, and the product was obtained by evaporation of the
2
2
allowed estimation of the relative yields of trans-pinocarveol (14)
and myrtenol (15). The data of the 1H NMR were in agreement
with those reported previously [17–20].
1
1
solvent and analysed by H NMR. The H NMR data were in agree-
ment with those reported previously [14,15].
2
.3.3. Ascaridole (8)
1
H NMR (300 MHz, CDCl3): δH = 0.97 (3H, d, J 6.90, CH3),
0
.98 (3H, d, J 6.9, CH3), 1.31 (3H, s, CH3), 1.51–1.56 (2H, m), 1.85
(H, sept, J 6.90, isopropyl), 1.97–1.92 (2H, m), 6.42 (H, d, J 8.58,
olefinic CH), 6.53 ppm (H, d, J 8.58, olefinic CH); MS (EI, 70 eV):
m/z 168 (M+, 1%), 150 (7%), 134 (32%), 119 (100%), 107 (33%), 91
2
.3.9. Trans-pinocarveol (14)
(37%).
1
H (300 MHz, CDCl3): δH = 0.64 (s, 3H, 9-CH3), 1.27 (s, 3H, 8-
CH3), 1.72 (d, 1H, J = 9.8 Hz, 7-H), 1.84 (dd, 1H, J = 14.6, 4.2 Hz,
2
.3.4. p-Cymene
4
2
1
1
7
-Hb); 1.99 (m, 1H, 5-H), 2.23 (m, 1H, 4-Ha), 2.37 (m, 1H, 7-H),
.51 (t, 1H, J = 5.5 Hz, 1-H); 4.42 (d, J = 7.6 Hz, 1H, 3-H), 4.82 (s,
H, 10-Ha), 5.00 (s, 1H, 10-Hb). MS (EI, 70 eV): m/z = 152 (M+,
%), 134 (33%), 119 (40%), 109 (30%), 91 (72%), 92 (100%), 83 (54%),
0 (52%).
1
·
H (300 MHz, CDCl3): δH = 1.22 (3H, d, J 1.68, CH ), 1.24 (3H,
3
·
·
3
d, J 1.68, CH ), 2.31 (3H, s, CH ), 2.87 (H, sept, isopropyl), 7.11
3
·
+
(4H, s, Ar-H ); MS (EI, 70 eV): m/z = 134 (M , 30%), 119 (100%),
115 (6%), 103 (4%), 91 (18%), 77 (5%).
2
.3.5. Citronellol (9)
The substrate (4.9 mmol) in 65 mL of chloroform was mixed
2.3.10. Myrtenol (15)
−4
1
with the appropriate amount of photosensitizer (9.8 × 10
mmol
H (300 MHz, CDCl3): δH = 0.83 (s, 3H, 9-CH3), 1.29 (s, 3H, 8-
of porphyrin or supported porphyrin) to originate a sensitizer-to-
substrate molar ratio of 1/5000. Then 203 mg of base (sodium
hydrogen carbonate) was added. GC was used to monitor the evo-
lution of the reaction after disappearance of the reagent. After
filtration of the photosensitizer and evaporation of the solvent,
CH3), 1.17 (d, 1H, J = 8.6 Hz, 7-H), 2.13 (m, 1H, 5-H); 2.13 (m, 1H,
1-H), 2.24 (m, 1H, 4-Ha), 2.27 (m, 1H, 4-Hb); 2.41 (m, 1H, 7-H),
3.98 (m, 2H, 10-H); 5.47 (m, 1H, 3-H). MS (EI, 70 eV): m/z = 152
(M+, 4%), 134 (1%), 119 (16%), 108 (31%), 91 (49%), 79 (100%).
The relative amounts of trans-pinocarveol (14) and myrtenol
1
1
the product was obtained and analysed by H NMR. The propor-
tion of the two regioisomers thus obtained, (E)-7-hydroperoxy-3,7-
(15) were estimated by H NMR: % for (14) = areaδ =4.42 ×
H
100/(areaδ =5.47 + areaδ =4.42).
H
H

S.M. Ribeiro et al. / Journal of Catalysis 256 (2008) 331–337
333
Scheme 1. PS1 synthesis sequence.
Scheme 2. PS2–PS4 synthesis sequence.
3
. Results and discussion
of porphyrin was carried out by controlled chlorosulphonation of
followed by the reaction of the corresponding chlorosulpho-
nyl derivatives (1S) with DAP, thus obtaining photosensitizer PS1
1
3
.1. Synthesis of supported photosensitizers
(
Scheme 1).
In our previous report, we established a simple route to co-
Because nonsymmetrical porphyrin 1 is obtained in a low yield,
valently link nonsymmetrical porphyrin 1 to a Merrifield polymer
structure [12]. The strategy requires the reaction of the Merri-
field polymer with an excess of 1,12-diaminododecane to obtain
a dodecyl aminopolymer (DAP) derivative. The covalent linkage
we tried to substitute it by symmetrical tetraarylporphyrins 2 to
4, which are obtained in much higher yields. Following the same
strategy as for chlorosulphonation, we obtained tetrachlorosulpho-
nyl derivatives (2S–4S) instead of monochlorosulphonyl deriva-

334
S.M. Ribeiro et al. / Journal of Catalysis 256 (2008) 331–337
Table 1
Values for porphyrin incorporation (mmol g 1) for photosensitizers PS1 to PS5
Table 2
−
Results for the photooxidation of α-terpinene with PS1 to PS5 and porphyrin 6
Polymer–spacer–porphyrin
Values of
Entry
porphyrin
incorporation
−1
(mmol g
)
a
8 (%)b
0
.04
Photosensitizer
R = nph/nterp
Time (h)
1
2
3
4
5
6
7
8
8
9
PS1
PS2
PS3
PS5
6
PS1
PS2
PS3
PS4
PS5
6
1/600
2.5
3
3
87 (13)
92 (8)
91 (9)
1/600
1/600
1/600
1/600
1/5000
1/5000
1/5000
1/5000
1/5000
1/5000
4
84 (16)
88 (12)
86 (14)
83 (17)
86 (14)
84 (16)
66 (34)
93 (7)
(
PS1)
1.5
3.5
4.5
7
11
8.2
2.3
0
.02
10
a
Photosensitizer/α-terpinene ratio.
%
b
of 8 in reaction mixture by 1H NMR. In parentheses the amount of p-cymene.
(
PS2)
would expect from the presence of more chlorosulphonyl groups.
With bromo derivative (2S), this loading actually was lower. Sur-
prisingly, PS5 exhibited the highest loading of the photosensitizers
tested. In the final washing of PS5, we isolated and identified by
mass spectrometry a porphyrin with one sulphonic group, certainly
due to hydrolysis of the chlorosulphonyl group of 1S. By infrared
spectroscopy, we noted that the band for the CH2Cl benzylic group
0
.05
(
(
(
PS3)
PS4)
PS5)
−1
(1264 cm ) was significantly decreased relative to the original
Merrifield polymer, suggesting some substitution at this position.
Washing with triethylamine in dichloromethane did not reduce
this loading value, suggesting a strong bond between porphyrin
and polymer. The comparison of the infrared absorption zones for
0
.04
−1
−1
the sulphonic groups [22], 1370–1330 cm
and 1200–1145 cm
of the Merrifield polymer and the isolated porphyrin with a sul-
phonic group, suggests that in this case, a direct bond of porphyrin
to polymer structure occurred from displacement of the benzylic
chlorine by the free sulphonic group.
0.20
3
.2. Photooxidation reactions
We began the evaluation of the supported photosensitizers
PS1–PS5 by attempting the photooxidation of α-terpinene (7) [23]
using chloroform as the solvent, air as the oxygen supply, and
sensitizer-to-substrate ratios of 1:600 and 1:5000. For supported
catalysts, the polymer was filtered at the end of the reaction and
tives as in the case of porphyrin 1. The reaction of these tetra-
chlorosulphonyl derivatives with a polymer may increase the like-
lihood of bonding with the aminogroups of DAP. Nonbonding
chlorosulphonyl groups possibly hydrolyze in the subsequent treat-
ments. Photosensitizers PS2–PS4 were prepared in this manner
the solvent evaporated. For free porphyrin TDCPP (6), a chromato-
graphic separation was needed. 1H NMR analysis of the reaction
residue showed that the main product was ascaridole (8), as was
confirmed by GC/MS analysis. Some p-cymene was detected by
(
Scheme 2).
GC/MS and quantified by 1H NMR. The results for PS1–PS5 and
We also attempted the reaction of 1S directly with the Merri-
TDCPP (6) are given in Table 2.
field polymer, originating photosensitizer PS5, which has no spacer
between the polymer structure and the porphyrin. We performed
comparative studies between our photosensitizers (PS1–PS5) and
tetra (2,6-dichlorophenyl) porphyrin (TDCPP) (6), one of the most
active sensitisers used in photooxidation [21].
Values for the amount of porphyrin bonded to DAP were calcu-
lated from the nitrogen content obtained by elemental analysis of
polymers PS1–PS4, discounting the value of nitrogen correspond-
ing to the initial DAP polymer. For PS5, the porphyrin content was
calculated directly from the nitrogen values (Table 1).
The results in Table 2 show that, compared with porphyrin 6,
the photosensitizers tested exhibited moderate to good activity
as singlet oxygen generators. A blank experiment with Merrifield
polymer demonstrated no product formation after 13 h of reaction.
The different photosensitizers exhibited much the same reactiv-
ity order, 6 > PS1 > PS2, PS3 > PS5 for the ratio of 1/600 and
6 > PS1 > PS2 > PS3 > PS5 > PS4 for the ratio of 1/5000. Pho-
tosensitizer PS5 had lower activity than PS1, PS2 and PS3, as was
expected based on the absence of the C12 chain spacer [12]; how-
ever, PS5 was more active than PS4, which had the C12 spacer.
In this case, we conjecture that the positive effect of the spacer
in PS4 was counter weighed by the absence of the ortho halogen
atoms in porphyrin structure. These halogens are important to pro-
Using the tetrachlorosulphonyl derivatives (2S–4S), we found
no increase in the loading of the porphyrins to the amino poly-
mer (DAP) compared with the monosulphonyl derivative 1S, as we

S.M. Ribeiro et al. / Journal of Catalysis 256 (2008) 331–337
335
Table 3
Table 5
Photooxidation of α-terpinene with PS1 or porphyrin 6 using different photosen-
sitizer/α-terpinene ratios
Photooxidation of α-terpinene in consecutive experiments with a photosensitizer/
α-terpinene ratio of 1/5000
a
8 (%)b
8 (%)a
TOFb
Entry
Photosensitizer
R = nph/nterp
Time (h)
Photosensitizer
Reaction
Time (h)
1
2
3
4
5
6
7
8
6
6
1/600
1.5
2.3
2.5
3.3
3.3
4
88 (12)
91 (9)
87 (13)
95 (5)
86 (14)
82 (18)
65 (35)
51 (49)
PS1
1st
2nd
3rd
3.3
3.3
2.5
86 (14)
86 (14)
87 (13)
1303
1303
1740
1/5000
1/600
PS1
PS1
PS1
PS1
PS1
PS1
1/2000
1/5000
1/15000
1/30000
1/60000
PS2
PS3
1st
4.5
5.1
5.1
83 (17)
86 (14)
84 (16)
922
843
823
2
3
nd
rd
10
19.3
1st
2nd
3rd
7
8.5
9
82 (18)
84 (16)
79 (21)
586
494
439
a
Photosensitizer/α-terpinene ratio.
%
b
of 8 in reaction mixture by NMR. In parentheses the amount of p-cymene.
PS5
1st
2nd
8
12
66 (34)
63 (37)
413
263
Table 4
Results for consecutive photooxidations of α-terpinene using photosensitizer PS1
a
of 8 in reaction mixture by 1H NMR. In parentheses figures the amount of
%
a
8 (%)b
TOFc
Reaction
R = nph/nterp
Time (h)
p-cymene.
b
Moles of 8/moles of photosensitizer × h.
1
2
3
1
2
3
1
2
3
st
1/600
2.5
2
87 (13)
86 (14)
85 (15)
86 (14)
86 (14)
87 (13)
82 (18)
78 (22)
81 (19)
209
258
nd
rd
st
1/600
1/600
2.5
3.3
3.3
2.5
4
204
1/5000
1/5000
1/5000
1/15000
1/15000
1/15000
1303
1303
1740
3075
1463
1429
nd
rd
st
nd
rd
8
8.5
a
PS1/α-terpinene ratio.
b
Scheme 3. Citronellol photooxidation products.
%
of 8 in reaction mixture by NMR. In parentheses figures the amount of p-cy-
mene.
c
Moles of 8/moles of PS1 × h.
1/15,000. At this photosensitizer-to-substrate ratio, a clear decrease
in photosensitizer activity was seen after the first reaction, dou-
bling the reaction time, but the activity for the third consecutive
reaction was maintained. This corresponds to 45,000 total reaction
cycles, similar to the value obtained with the Griesbeck photoox-
idation system [16]. With simple filtration of the photosensitizer,
about 4.6 g of ascaridole (8) could be obtained using only 35 mg
of PS1 after 3 consecutive reactions.
From a practical standpoint, photosensitizers PS2, PS3, and PS5
can be prepared more easily than PS1 and, if sufficiently active, can
be considered better solutions for large-scale processes. Table 5
compares the activity of PS1, PS2, PS3, and PS5 in α-terpinene
oxidation and in consecutive reactions. Supported photosensitiz-
ers with a C12 carbon chain spacer (PS1–PS3) exhibited almost the
same activity during the three reaction cycles. In contrast, the cat-
alyst without the spacer, PS5, was the least active, exhibiting a
slower second reaction cycle and a greater amount of p-cymene.
As indicated by its TOF values, PS1 was the most active photo-
sensitizer, with activity superior to that of PS2 and PS3. The pho-
tosensitizer with bromine groups, PS2, seemed to be much more
active than PS3 with chlorine atoms. The importance of the spacer
can be demonstrated by comparing the results for PS1 and PS5.
PS1 has a porphyrin with similar structure to PS5, spaced from
the polymer backbone by a C12 chain, and is a much more active
photosensitizer.
To further investigate the scope of this type of supported pho-
tosensitizer, we carried out reactions using other monoterpenes as
substrates. Photooxidation of citronellol (9) originated the isomeric
hydroperoxides (10) and (11) [25–30], which, after reduction and
acid cyclization [31], could be converted to a mixture of cis and
trans rose oxides, which are valuable compounds in the perfume
industry (Scheme 3).
With the more active photosensitizers PS1 and PS2 and free
porphyrin 6, we attempted photooxidation of citronellol (9) us-
ing different photosensitizer-to-substrate ratios. The results, given
in Table 6, show that citronellol was more difficult to oxidize
than α-terpinene, indicating that supported photosensitizers had
much slower reactions than free porphyrin but gave the same iso-
mote stability of the macrocycle [21] and to activate singlet oxygen
generation by the heavy atom effect [24]. The supported catalyst
with the spacer and the ortho chlorine atoms, PS1, proved to be
the most active, with a performance close to that of the nonsup-
ported catalyst 6. Crude reaction products showed a mass balance
corresponding to near quantitative conversions. NMR analysis re-
vealed ascaridole (8) as the main product, with varying amounts
of p-cymene due to ring oxidation of α-terpinene. Table 2 shows
that for the photooxidation reaction, selectivity was higher at the
1
/600 ratio than at the 1/5000 ratio. Considering that commer-
1
cial α-terpinene contains about 5% p-cymene (by H NMR), the
supported photosensitizers generally exhibited very good selectiv-
ity for the formation of ascaridole 8. Only in the case of PS5, which
does not have the C12 chain spacer, and at the 1/5000 ratio did we
obtain an exceptional amount of p-cymene.
Of all photosensitizers tested, PS1 showed the best activity. To
analyse its photocatalytic ability, we carried out the photooxida-
tion of α-terpinene by increasing the photosensitizer-to-substrate
ratio up to 1/60,000 (Table 3). For comparison, we also carried out
some reactions with porphyrin 6 with this substrate. The results
demonstrate that at photosensitizer-to-α-terpinene ratios of up to
1
/15,000, PS1 showed good catalytic activity, originating mainly
the product from singlet oxygen reaction. At a ratio of 1/30,000,
the reaction was much slower, and the amount of p-cymene in-
creased considerably. At a ratio of 1/60,000 ratio, the yields of the
oxygenated product and p-cymene were equivalent.
We further studied the efficiency of our supported catalysts af-
ter recovery and recycling. We evaluated this efficiency for PS1 in
α-terpinene photooxidation at three different catalyst-to-substrate
ratios. The results, given in Table 4, show that for photosensitizer-
to-substrate ratios of 1/600 and 1/5000, PS1 was able to carry
out 3 consecutive reactions with no significant loss of activity
and with good selectivity for ascaridole (8). Turnover frequencies
(TOFs) remained the same in the sequential cycles, indicating good
catalyst stability. TOF values were higher at higher photosensitizer-
to-substrate ratios, reaching a maximum of 3075 at a ratio of

336
S.M. Ribeiro et al. / Journal of Catalysis 256 (2008) 331–337
Scheme 4. α-Pinene and β-pinene photooxidation products.
Scheme 5. Proposed α-pinene and β-pinene photooxidation reaction mechanism.
Table 6
Table 7
Results of citronellol photooxidation using supported photosensitizers PS1 and PS2
and free porphyrin 6
Results of α-pinene and β-pinene photooxidations using supported photosensitizers
PS1 and PS2 and free porphyrins 4 and 6. Photosensitizer/substrate ratio of 1/5000
Photosen- Photosensitizer/ Reaction
Isolated
yields (%)
Product distribution (%)a
10)
Substrate
Entry
Photosen-
sitizer
Time
(h)
Products
yields (%)
14
(%)
15
(%)
a
b
b
sitizer
substrate ratio
time (h)
(
(11)
1
2
3
4
5
6
4
PS1
PS2
PS1c
20
18
24
88
29
54
65
85
55
64
42
88
69
46
81
58
12
31
54
19
6
6
PS1
PS1
PS2
PS2
1/600
1/5000
1/600
1/5000
1/600
1/5000
1.5
4
6.5
10
8
99
99
98
99
99
99
49
50
47
47
48
45
51
50
53
53
52
55
25
6
7
8
6
PS1
PS2
24
37
138
59
53
53
34
32
26
66
68
74
a
Calculated from 1H NMR of the crude reaction mixture.
lated yields. Among the supported photosensitizers, PS1 again was
more active than PS2, particularly when higher amounts of sub-
strate relative to catalyst were used. The results in Table 6 indicate
a slight excess of allylic hydroperoxide (11) in PS1 and PS2, in
contrast to other systems in which allylic hydroperoxide (10) is
favoured [16,29].
a
Isolated yields after reduction.
b
c
Estimated from 1H NMR of the isolated product (Section 2).
Photosensitizer/substrate ratio of 1/600.
pinene and, as expected, β-pinene was less reactive than α-pi-
nene [33]. In both cases, relative to the other substrates, there were
some differences considering the formation of different oxidation
products. The product expected for the singlet oxygen addition to
12 was trans-pinocarveol 14, and that expected for this addition to
Photooxidation of α-pinene (12) and β-pinene (13) with PS1,
PS2, and 6, followed by reduction of the corresponding hydroper-
oxide products with triphenylphosphine [32] and chromatographic
isolation, gave trans-pinocarveol (14) and another product, identi-
1
13
fied by comparison of the H and C spectra as myrtenol (15) [17–
1
3 myrtenol was 15, both of which originated from an ene reac-
20] (Scheme 4, Table 7).
The results demonstrate that with these latter two monoter-
penes, reactions were slower and lower product yields were ob-
tained compared with those obtained with citronellol and α-ter-
tion [34]. We also observed products coming from non-ene origin,
15 in the case of α-pinene and 14 in the case of β-pinene. In some
cases, these were the main products. A possible explanation for

S.M. Ribeiro et al. / Journal of Catalysis 256 (2008) 331–337
337
the appearance of these unexpected products from substrates like
References
α-pinene and β-pinene is a mechanism of electron transfer be-
tween excited photosensitizers and these substrates, as proposed
by Zhang [33] (Scheme 5).
[1] E.L. Clennan, A. Pace, Tetrahedron 61 (2005) 6665.
[2] M.C. DeRosa, R.J. Crutchley, Coord. Chem. Rev. 233-234 (2002) 351.
[
[
[
3] I. Margaros, T. Montagnon, M. Tofi, E. Pavlakos, G. Vassilikogiannakis, Tetrahe-
dron 62 (2005) 5308.
4] J. Wahlen, D. De Vos, P.A. Jacobs, P.L. Alsters, Adv. Synth. Catal. 346 (2004)
The electron transfer between 12 or 13 and the excited pho-
tosensitizer gave the same radical cation 16 and the radical anion
photosensitizer, which can mediate the ene reaction by a super-
oxide radical via type I photooxygenation (pathway B) [35]. Alter-
natively, species 16 can lose a proton to the hydrogen carbonate,
giving the allyl radical 17, which, by a radical chain reaction with
triplet oxygen, leads to the hydroperoxides 18 and 19. This cor-
responds to the non-ene reaction (pathway A) a route supported
by some data. Reaction of PS1 and α-pinene at a low substrate
ratio (1/600, entry 5) gave more product from ene addition than
the reaction with a higher substrate ratio (1/5000, entry 3). Be-
cause electron-transfer processes require proximity between the
donor and acceptor species, decreasing the amount of substrate
molecules decreases the likelihood of the electron-transfer process.
Reaction with porphyrin 4, which must have a more negative re-
duction potential than porphyrin 6 [36], disfavoured pathway A,
and thus the amount of non-ene products would be expected to
be lower, as was seen (entry 2). More compelling evidence that
pathway A is operative in these reactions came from the photoox-
idation of α-pinene with PS1 in the presence of BHT, a known
radical scavenger. The reaction was much slower (64 h) because
pathway A was inoperative. After reduction and chromatographic
workup, no myrtenol was detected; only the trans-pinocarveol was
isolated.
152.
5] K. Gollnick, Adv. Photochem. 8 (1968) 1.
[6] M. Benaglia, T. Danelli, F. Fabris, D. Sperandio, G. Pozzi, Org. Lett. 4 (2002)
4229.
[7] A.G. Griesbeck, T.T. El-Idreesy, Photochem. Photobiol. Sci. 4 (2005) 205.
[8] A.G. Griesbeck, T.T. El-Idreesy, J. Lex, Tetrahedron 62 (2006) 10615.
[9] J.J. Inbaraj, M.V. Vinodu, R. Gandhidasan, R. Murugesan, M. Padmanabhan,
J. Appl. Polym. Sci. 89 (2003) 3925.
[
10] C.J. Rogers, T.J. Dickerson, P. Wentworth Jr., K.D. Janda, Tetrahedron 61 (2005)
12140.
[11] N. Kitamura, K. Yamada, K. Ueno, S. Iwata, J. Photochem. Photobiol.
Chem. 184 (2006) 170.
12] S.M. Ribeiro, A.C. Serra, A.M.d’A. Rocha Gonsalves, Tetrahedron 63 (2007) 7885.
13] R.A. Johnstone, M.L.P.G. Nunes, M.M. Pereira, A.M.d’A. Rocha Gonsalves, A.C.
Serra, Heterocycles 43 (1996) 1423.
14] H. Tokuyama, E. Nakamura, J. Org. Chem. 59 (1994) 113.
[15] C.J. Pouchert, in: The Aldrich Library of NMR Spectra, vol. 1, second ed., Aldrich
Chemical, Millwaukee, WI, 1983, p. 742.
16] A.G. Griesbeck, T.T. El-Idreesy, A. Bartoschek, Adv. Synth. Catal. 346 (2004) 245.
17] E. Okuyama, K. Umeyama, Y. Saito, M. Yamazaki, M. Satake, Chem. Pharm.
Bull. 41 (1993) 1309.
A
[
[
[
[
[
[
[
18] A.Y. Baddjah-Hadj-Ahmed, B.Y. Meklati, Magn. Reson. Chem. 30 (1992) 807.
19] S.-G. Lee, Magn. Reson. Chem. 40 (2002) 311.
[20] C. Filliatre, J.J. Villenave, J. Prévot, Bull. Soc. Chim. Fr. 9–10 (1979) II-473.
[21] H. Quast, T. Dietz, A. Witzel, Liebigs Ann. (1995) 1495.
[
22] D.H. Williams, I. Fleming, Spectroscopy Methods in Organic Chemistry, third
ed., McGraw–Hill, London, 1980, p. 64.
23] G.O. Schenck, Angew. Chem. 69 (1975) 579.
4
. Conclusion
[
[
24] E.G. Azenha, A.C. Serra, M. Pineiro, M.M. Pereira, J. Seixas de Melo, L.G. Arnaut,
S.J. Formosinho, A.M.d‘A. Rocha Gonsalves, Chem. Phys. 280 (2002) 177.
25] R. Gerdes, O. Bartels, G. Schneider, D. Wohrle, G. Schulz-Ekloff, Polym. Adv.
Technol. 12 (2001) 152.
Heterogeneous catalysts for singlet oxygen generation were pre-
pared from the chlorosulphonation of porphyrins and reaction
with amino alkylated Merrifield resin in a very simplified prepa-
ration procedure that compares favourably with those reported
previously. The sensitizer with chlorine atoms in the porphyrin
structure PS1 proved to be very active for the photooxidation
of α-terpinene and citronellol, demonstrating efficiency in higher
substrate-to-catalyst ratios, up to 1/30,000. With this catalyst, con-
secutive reactions could achieve up to 45,000 oxidation cycles. In
the case of poorly reactive substrates (i.e., α-pinene and β-pinene),
non-ene products were formed through a pathway involving an
electron-transfer mechanism.
[
[26] S.A. Chavan, W. Maes, L.E.M. Gevers, J. Wahlen, I.F.J. Vankelecom, P.A. Jacobs,
W. Dehaen, D.E. De Vos, Chem. Eur. J. 11 (2005) 6754.
[
27] M. Oedlgemoller, C. Jung, J. Mattay, E. Zimmermann, Green Chem. 7 (2005)
35.
[28] T. Hino, T. Anzai, N. Kuramoto, Tetrahedron Lett. 47 (2006) 1429.
[29] M.J. Fuchter, B.M. Hoffman, A.G.M. Barrett, J. Org. Chem. 71 (2006) 724.
[30] S. Meyer, D. Tietze, S. Rau, B. Schafer, G. Kreisel, J. Photochem. Photobiol. A 186
(2007) 248.
[31] W. Pickenhagen, D. Schatkowski, US Patent 5 892 059 (1999), to Dragoco Ger-
berding & Co. Aktiengesellschaft.
[32] W. Adam, B. Nestler, Liebigs Ann. Chem. (1990) 1051.
[33] B.-W. Zhang, Y.-F. Ming, Y. Cao, Photochem. Photobiol. 40 (1984) 581.
[34] M. Prein, W. Adam, Angew. Chem. Int. Ed. Engl. 35 (1996) 477.
[35] C.S. Foote, Photochem. Photobiol. 54 (1991) 659.
[36] A. Girandeau, H.J. Callot, J. Jordan, I. Ezhar, M. Gross, J. Am. Chem. Soc. 101
(1979) 3857.
Acknowledgment
The authors thank Chymiotechnon and FCT-POCTI/QUI/55931/
004 for financial support.
2