Selective epoxidation of unsaturated fatty esters over peroxophosphotungstic catalysts (POW) under solvent free conditions: Study of the POW catalyst's mechanism

Article abstract of DOI:10.1016/j.cattod.2010.02.005

Methyl oleate epoxidation with a peroxophosphotungstate (POW) catalyst was performed under solvent free conditions. After underlying a synergistic effect of hydrogen peroxide and the dioxygen flow to obtain a total epoxide yield, new investigations were realized to understand this effect. The POW catalyst mechanism was studied using labeled oxygen flow and radical traps. We concluded that the reaction occurs in two parallel ways, a major one thanks to a concerted mechanism and a minor radicalar one leading to the formation of methyl stearate alcohol and diol intermediates. Oxygen flow is only necessary to perform the radicalar way.

Full text of DOI:10.1016/j.cattod.2010.02.005

Catalysis Today 157 (2010) 371–377
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Selective epoxidation of unsaturated fatty esters over peroxophosphotungstic
catalysts (POW) under solvent free conditions: Study of the POW catalyst’s
mechanism
Evelyne Polia, Nicolas Biona, Joël Barraulta, Stefano Casciatob, Vincent Duboisb,
Yannick Pouillouxa, Jean-Marc Clacensa,∗
a Université de Poitiers, Laboratoire de Catalyse en Chimie Organique, CNRS, UMR 6503, ESIP, 40, avenue du Recteur Pineau, 86022 Poitiers Cedex, France
b Institut Meurice, Unité de Chimie Physique et Catalyse, Avenue E. Gryson, 1 Bât 10, 1070 Bruxelles, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 6 March 2010
Methyl oleate epoxidation with a peroxophosphotungstate (POW) catalyst was performed under solvent
free conditions. After underlying a synergistic effect of hydrogen peroxide and the dioxygen flow to
obtain a total epoxide yield, new investigations were realized to understand this effect. The POW catalyst
mechanism was studied using labeled oxygen flow and radical traps. We concluded that the reaction
occurs in two parallel ways, a major one thanks to a concerted mechanism and a minor radicalar one
leading to the formation of methyl stearate alcohol and diol intermediates. Oxygen flow is only necessary
to perform the radicalar way.
Keywords:
Fatty acid methyl esters
Selective epoxidation
Green oxidation
H2O2
© 2010 Elsevier B.V. All rights reserved.
Peroxophosphotungstate catalyst
1. Introduction
low temperature uses). On the other hand, the MO epoxide is more
stable than the unsaturated fatty methyl starting ester.
Due to the increase of the greenhouse effect and the reduc-
tion of the petroleum oil resources, renewable oils and fats of
vegetable origin is an important issue in chemistry since a shift
from petrochemical feedstocks to renewable resources can con-
tribute to a sustainable development. Vegetable oils such as
rapeseed or sunflower oils are agroresources that can produce var-
ious compounds by hydrolysis, methanolysis, . . . To extend the
large scale potential of agroresources, new and green transfor-
mation of fatty acids must be developed. Numerous studies were
focused on the carboxylic function transformation but it is also
interesting to develop green process for the functionalization of
the fatty alkyl chain and particularly the unsaturated function.
Oleic acid is a mono unsaturated fatty acid which can be pro-
duced from a hybrid sunflower called “oleic sunflower” giving
between 80 and 95% of oleic acid, is one of the most abundant one
[1].
Industrially the epoxidation of fatty esters uses a mixture of
formic acid/hydrogen peroxide or peracids [4]. This process induces
the formation of various by-products, and solvents elimination is
needed. That is expensive and there are explosion risks.
The direct epoxidation of unsaturated fatty compounds by H2O2
as oxidant was investigated since H2O2 is a good candidate to
develop a green process. It is cheap, readily available and trans-
formed into water as the only by-product. Moreover, that is easy
to eliminate by decantation because of the biphasic system. During
the last decade, many different catalytic systems for the epoxida-
tion using hydrogen peroxide have been studied [5–7]. Catalysts
based on titanium [8], manganese [9], tungsten [10] and rhenium
[11] have been described.
A peroxophosphotungstate (POW) based catalyst was suc-
cessfully developed by Venturello and D’Aloisio [12] to perform
epoxidation reaction in a biphasic media containing an organic sol-
vent. Because many side reactions occur (hydrogen transfer, double
bond shift, cis–trans isomerization, branching), a complex prod-
uct mixture could be obtained. During the epoxidation of FAMEs,
oligomerization by-products [13] can be formed; the main prod-
ucts are dimers, trimers, and isostearic acid (Scheme 1). Residual
tungstic acid coming from the POW catalyst could also lead to
another by-product formation, the 9,10-dihydroxystearic methyl
ester as previously described by Santacesaria et al. [14].
In this general context, we have focused our study on the syn-
thesis of fatty epoxides from methyl oleate (MO) which are an
important building block compound [2]. Indeed, MO epoxide is
an intermediate of the polymers chemistry (such as the class of
polyurethanes, polycyanates), in the lubricants and detergents syn-
thesis, and in the formulation of biofuels [3] (as an additive for
During a first set of experiments [15], reaction parameters of
the methyl oleate epoxidation reaction with a peroxophospho-
∗
Corresponding author. Tel.: +33 549454417; fax: +33 549453349.
E-mail address: jean-marc.clacens@univ-poitiers.fr (J.-M. Clacens).
0920-5861/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.cattod.2010.02.005

372
E. Poli et al. / Catalysis Today 157 (2010) 371–377
Scheme 1. Main by-products formed during the oligomerization of FAMEs.
tungstate catalyst, hydrogen peroxide as oxidant and under solvent
free condition were studied. After 0.5 h of reaction time, at 313 K
and with a molar ratio MO/H2O2 (1/3), 97.4% of MO conversion
and 73.3% of epoxide yield were obtained. To keep this yield while
decreasing the amount of the oxidant (hydrogen peroxide) to an
equimolar ratio MO/H2O2, the reaction was performed with a gas
flow.
using the homogeneous POW catalyst under the above-mentioned
reaction conditions.
2. Experimental
2.1. Preparation of the catalyst
The nature of the reaction atmosphere was then studied using
different carrier gases (Table 1) with an equimolar ratio of the reac-
tants. By replacing the air atmosphere (no flow), by an air flow
bubbling in the reaction media, the epoxide yield was strongly
increased from 71.4% to 95.0%. Moreover, we evidenced a syn-
ergistic effect when using both air and hydrogen peroxide, since
after the replacement of air by nitrogen, a drastic decrease of the
epoxide yield from 95% to 56% was obtained. If the reaction was
carried out under an O2 flow, 100% of epoxide selectivity was
obtained with an equimolar amount of oxidant under greener con-
ditions. These results confirmed the importance of the presence of
O2 in the epoxidation reaction atmosphere. Nevertheless, the use
of an O2 flow cannot replace the use of H2O2 as oxidant. Indeed,
if the reaction was carried out without H2O2, no conversion was
obtained.
The peroxophosphotungstate catalyst ([(C8H17)3NCH3]3
[PO4[W(O)(O2)2]4]) was prepared according to Venturello and
D’Aloisio [12] and analyzed by IR for the confirmation of its
structure. In a typical procedure a suspension of 5 g of tungstic acid
(Fluka) in 14 ml of H2O2 35% (w/v) (Acros Organics) was stirred
and heated at 333 K until a colorless solution was obtained. After
filtration and cooling the solution at room temperature, 1.24 ml of
H3PO4 40% (w/v) (85%, v/v Prolabo) were added and this solution
was diluted with 60 ml of distilled water. To the resultant solution,
4.18 g of methyltrioctylammonium chloride (Fluka) in 80 ml of
dichoromethane (Carlo-Erba) was added dropwise (2 min) under
strong stirring. Then the mixture was stirred about 15 min. After
decantation, the organic phase was dried over MgSO4 (SDS) and
gently evaporated on a rotary evaporator under reduced pressure
at 308–313 K to obtain slightly yellow syrup.
Noyori et al. [16] proposed a concerted mechanism, based on a
labeled H2O2 study, to explain this catalyst working but Noyori did
not study the influence of oxygen flow on the epoxidation reaction.
The present study is then dedicated to the understanding of
the O2 flow effect as well as the proposal of a reaction mechanism
The ruthenium catalyst was prepared following the incipi-
ent wetness impregnation. This method consists in blending the
support with a volume of ruthenium chloride (RuCl3) solution cor-
responding to volume pore of the support. The concentration of the
solution is adjusted for obtaining the desired amount of ruthenium
(5 wt%) in the final catalyst. After the impregnation, the catalyst is
dried at 423 K under nitrogen.
Table 1
Effect of atmosphere gas phase on the reaction conversion and selectivity under the
following conditions: H2O2/MO (1/1), 313 K, 30 min.
2.2. Catalytic test
Conditions
MO conversion (%)
MO epoxide selectivity (%)
In a double-walled reactor (with internal diameter of 16 mm)
the peroxophosphotungstate catalyst syrup was exactly weighted
directly into the reactor (0.094 mmol), then the methyl oleate (MO)
(99% Aldrich, 3.37 mmol) was added under strong stirring. When a
temperature of 313 K was reached, H2O2 35% (w/v) (330 ␮l) were
N2 (200 ml/min)
Air (no flow)
Air (200 ml/min)
O2 (50 ml/min)
83.0
97.4
97.8
99.1
81.8
75.3
99.3
98.4

E. Poli et al. / Catalysis Today 157 (2010) 371–377
373
added. After 30 min of reaction, 25 ml of water (room tempera-
ture) were added in order to stop the reaction. The organic phase
was then recovered and extracted with 20 ml of ethyl acetate (99%
Carlo-Erba) and dried on MgSO4. After evaporation of the solvent,
the products were obtained and analyzed by GC.
Catalytic test in the presence of labeled 18O2 molecule was
carried out using a recycle closed system coupled to a mass spec-
trometer via a calibrated leak valve. This system was elsewhere
detailed for oxygen mobility measurement on oxides [17]. The
double-walled reactor described above was adapted to the system
to perform the epoxidation reaction following the same proto-
col but replacing air by 18O2/N2 atmosphere. Total volume of the
system was around 150 cm3 and recycle 18O2/N2 flow rate was
170 cm3 s−1. All the volume was purged with nitrogen before intro-
ducing labeled oxygen in order to prevent the presence of 16O2 in
the gas phase. The variation of oxygen isotopomer partial pressures
18O2 (m/z = 36), 18O16O (m/z = 34) and 16O2 (m/z = 32) as well as N2
(m/z = 28) were monitored every 40 s.
Fig. 1. The MO conversion and its epoxide yield at short reaction time (O2, 313 K,
H2O2/MO (1/1)).
The experiments done in the presence of radical scavengers
(0.2 mmol) were done under the same experimental conditions.
2.3.2. LC analysis
All the fatty compounds were analyzed using a Waters HPLC 600
Controller equipped with a Waters UV 486 Tunable Absorbance
Detector and a Waters 600 Pump. The column was a Nucleosil
C18 (250 mm × 4.6 mm × 5 ␮m) from Supelco and an autosampler
Waters 717 plus was used. The analyses were done at room tem-
perature and the solvent was methanol (Aldrich HPLC grade) with
a flow of 1 ml/min. The wavelength used to detect the fatty com-
pounds was 205 nm.
2.3. Analysis
2.3.1. GC analysis
All the compounds were analyzed using a Varian 3350 GC
equipped with an FID detector and an on-column injector. An HT5
column (25 m × 0.32 mm × 0.1 ␮m) from SGE was used. The injec-
tor and detector temperatures were respectively of 323 and 573 K.
The carrier gas was nitrogen. For analyte separation, the GC oven
temperature was fixed at 343 K for 1 min, then ramped at a rate
of 10 K/min to 473 K, and ramped at 25 K/min to 573 K then kept
constant for 1 min.
Before analysis, in a pillbox weight 0.1 g of the mixture product,
0.05 g of the external standard (dodecane) and 800 ␮l of solvent
(ethyl acetate) were added. After stirring, the sample was injected
into GC (0.1 ␮l).
3. Results and discussion
In a first time a mechanistic study was performed by analy-
sis of the reaction media at short reaction time under O2 flow to
determine what are the primary and secondary products. After,
to evaluate the effect of oxygen on the reaction and particularly
its influence on radical mechanism, some radical scavengers were
added at the beginning of the reaction. Finally, labeled 18O2 was
used to determine how the oxygen flow is involved in the epoxida-
tion reaction.
The methyl oleate conversion is expressed as follows at t time:
(Soi leate/Sdi od) − (Sot leate/Sdt od
)
Conversion (%) =
× 100
(Soi leate/Sdi od
)
3.1. Study of the MO conversion and selectivity at short reaction
time
where Soi leate: surface of the methyl oleate chromatographic peak
at time zero; Sdi od: surface of the dodecane chromatographic peak
at time zero; Sot leate: surface of the methyl oleate chromatographic
peak at time t; Sdt od: surface of the dodecane chromatographic peak
at time t.
The epoxidation reaction was performed under O2 flow at lower
reaction time to prove that the maximum of MO epoxide selectivity
is obtained at 0.5 h of reaction time and to determine the catalyst
activity at the early stage of the reaction.
The epoxide yield is expressed as follows at t time:
First of all, we can observe in Fig. 1 that the peroxophospho-
tungstate catalyst is very effective in the very first minutes of the
reaction. Indeed, it can be underlined that after only 5 min of reac-
tion; the conversion of the MO determined by GC analysis is already
of 32% and the epoxide yield of 13%. Furthermore, between 5 and
10 min of reaction time, the MO conversion is multiplied by 3 from
32% to 87%. At 10 min of reaction time the MO conversion is very
high (87%) but the MO epoxide yield is only 53%; so the carbon bal-
ance is uncompleted. Nevertheless, after 30 min of reaction time,
the conversion of the MO reaches 99% and the MO epoxide yield
reaches 99%. Only the peaks of MO and MO epoxide are identified
by GC analysis and it is impossible, using this technique, to deter-
mine the by-products of the MO conversion that would complete
the carbon balance.
(Setpox/St ) × Koleate
dod
Yield (%) =
× 100
(Soi leate/Sdi od) × Kepox
where Setpox: surface of the epoxide chromatographic peak at time
t; Sdt od: surface of the dodecane chromatographic peak at time t;
Soi leate: surface of the methyl oleate chromatographic peak at time
zero; Sdi od: surface of the dodecane chromatographic peak at time
zero; Koleate: relative response factor of the MO compared with
dodecane; Kepox: relative response of the epoxide compared with
dodecane.
The epoxide selectivity is expressed as follows:
yield (%)
conversion (%)
Selectivity (%) =
× 100
In order to determine the presence of GC undetected intermedi-
ate products, all the mixture samples were analyzed by HPLC with
a C18 Nucleosil column (Fig. 2).
After 5 min of reaction, the LC analysis shows the main pres-
ence of MO and the presence of MO epoxide. Two other products
The carbon balance is calculated from the initial mole of reac-
tants from GC analysis. All the yields were determined thanks to
calibrated GC peaks area.

374
E. Poli et al. / Catalysis Today 157 (2010) 371–377
Fig. 2. The MO epoxidation between 5 and 25 min of reaction time. HPLC analysis of the mixtures.
at 4.18 and 4.50 min (retention time) were identified by LCMS (not
shown) as the mono and diol of the MO, respectively methyl-10-
hydroxyoctadecanoate and methyl-9,10-dihydroxyoctadecanoate.
The two others peaks at 1.61 and 1.91 min of retention time were
attributed by LC to the organic cation of the homogeneous POW
catalyst; as confirmed by the HPLC analysis of the catalyst alone.
Fig. 3 shows the variation of the amount (based on peaks area) of
the MO, MO epoxide and MO diols during the epoxidation reaction
from the HPLC analysis presented previously in Fig. 2.
At 10 and 15 min of reaction time, the two relative amounts of
the mono and diol of the MO decreases while the MO epoxide rela-
tive amount increases. At 20 min of reaction time, there is still rise of
the MO epoxide relative amount and a decrease of the MO amount.
The mono and diol of MO have disappeared. Finally at 25 min of
reaction, MO is nearly fully converted into MO epoxide.
Finally, when the reaction time increases, it seems that first,
mono and diol of MO are formed, but these products decrease lead-
ing to an increase of the MO epoxide (Figs. 1 and 3). These data seem
indicate that MO mono and diols are reaction intermediates under
our reaction conditions. The diol is then transformed into epoxide
by dehydration because of the reaction media acidity (pH 1); it was
indeed confirmed from the pure diol which was fully transformed
into epoxide under our reaction conditions. The intermediate alco-
hol is also dehydrated to go back to the MO (Scheme 2). However
this observation is not in agreement with the catalytic mecha-
nism demonstrated by Noyori et al. [16], but this work concerned
the epoxide formation of less lipophilic molecules by a concerted
mechanism, in that case no alcohols were detected.
The formation of these intermediates could come from a rad-
icalar decomposition of H2O2 or from the reaction of MO with
tungstic acid as described by Santacesaria et al. [14] but under other
reaction conditions (70 ◦C, H2O2 excess, 6 h of reaction time). We
verified this last possibility, but under our reaction conditions, no
Fig. 3. Variation of the MO, MO epoxide and MO diols amounts (HPLC analysis).

E. Poli et al. / Catalysis Today 157 (2010) 371–377
375
Scheme 2. Reaction scheme proposal for the epoxidation of FAMEs by H2O2 with POW catalyst.
diol formation was observed using tungstic acid as catalyst.
of MO epoxide. Therefore, the epoxidation reaction does not involve
hydrogen donor and all sorts of radicals.
We can think that the epoxidation reaction with peroxophos-
photungstate and hydrogen peroxide as oxidant does not involve a
radical mechanism, as Noyori et al. [16] already explained.
However, we can mention that the use of these traps slows down
the reaction as the MO is not completely converted after 30 min of
reaction time.
When the MO epoxidation was performed with ascorbic acid as
radical trap (Fig. 4), no mono and diols of the MO were obtained at
short reaction time (10 min). In fact, it seems that there is a radical
initiation for the alcohols formation and that two parallels mech-
anisms during the epoxidation reaction occurs (Scheme 2); one
minor one that involve a radicalar formation of mono and diols,
and a major one that follow a concerted mechanism as described
by Noyori et al. [16].
The radicalar formation of mono and diols only occurs under
O2 or air flow while no alcohols were detected under a N2 flow
(not shown) meaning that the radicalar mechanism pathway does
not occur under N2. This is the reason why the reaction is slow
down under a nitrogen flow. So O2 only increase the reaction
rate.
3.2. Reaction in the presence of radical scavengers
To know if the epoxidation reaction with a peroxophospho-
tungstate catalyst and hydrogen peroxide is occurring through a
radical mechanism, the reaction was performed in the presence of
radical scavengers (0.2 mmol). If the catalyst mechanism involves
a radical mechanism in the presence of O2, the use of a radical trap
during the catalytic test, could prevent the MO epoxide formation.
The different traps used were histidine, BHT (2,6-di-ter-butyl-4-
methylphenol), and ascorbic acid. The histidine is a radical trap for
1O2 radicals; the BHT is a hydrogen donor and a radical trap for all
kind of radicals. Finally, ascorbic acid is a hydrogen donor, a trap of
1O2 and 3O2 radicals and a trap of all sorts of radicals. Fig. 4 shows
the results of MO epoxidation in the presence of each kind of radical
traps.
The addition of the histidine during the MO epoxidation does
not prevent the MO epoxide formation. This catalytic test proves
that the epoxidation reaction thanks to a peroxophosphotungstate
catalyst does not involve 1O2 radicals. The same reaction in the
presence of BHT and ascorbic acid does not prevent the formation
Fig. 4. MO epoxidation reactions in the presence of various radical scavengers (a, BHT; b, ascorbic acid; c, histidine: 0.5 h of reaction time; d, ascorbic acid: 10 min of reaction
time).

376
E. Poli et al. / Catalysis Today 157 (2010) 371–377
Fig. 5. Evolution of the gas phase during the MO epoxidation reaction in the pres-
Fig. 6. Evolution of the gas phase during the MO epoxidation reaction in the pres-
ence of H2O2, POW and 50 mbar of 18O2 in N2.
ence of a ruthenium based catalyst and 500 mbar of 18O2 in N2.
3.3. Effect of O2 by using 18O2
Taking into account the results presented previously with the
effect of an O2 flow, the epoxidation reaction was performed with
labeled oxygen to understand the role of O2 flow on the reaction
mechanism. By using 18O2/N2 flow in the catalytic test and analyz-
ing the products by GC–MS, it could be possible to discriminate the
origin of the oxygen of the epoxide formed. A consumption of 18O2
from the gas phase could have two reasons: (i) a consumption due
to the epoxidation reaction. In that case, the presence of 18O must
be found in the oxirane cycle of the epoxide molecule which could
be determined by GC/MS. (ii) A consumption due to the exchange
between oxygen from the gas phase and oxygenated molecules
of the reactants or the catalyst (POW, H2O2, MO for instance). In
that last case the disappearance of 18O2 should match with appear-
ance of 18O16O and/or 16O2 partial pressures in the gas phase. To
check this hypothesis, evolution of all the oxygen isotopomers were
collected on-line.
Fig. 7. GC–MS analysis of the MO epoxide obtained with the POW catalyst under a
18O2 flow.
Fig. 5 shows the gas composition during the MO epoxidation in
the presence of H2O2, POW as catalyst and under 18O2/N2 recycling
flow. After 30 min of reaction MO was totally converted whereas
18O2 partial pressure was kept constant and thus no other iso-
topomer of oxygen appeared in the gas phase. To confirm the
reliability of this technique, a ruthenium-based catalyst was tested.
This catalyst is known to be active for the epoxidation reaction by
using only O2 as oxidizing agent [18].
lar ion is at m/z = 157 instead of 155 corresponding to the presence
of labeled oxygen isotope. The same observation is done for the
second characteristic radical ion at m/z = 201 (199 + 2).
In the case of the POW catalyst, no labeled oxygen is consumed
during the epoxidation reaction, but as it was previously shown, the
epoxidation reaction reached total yield thanks to an O2 flow. So, if
under an O2 flow no epoxide were formed, but the MO epoxidation
was still performed, what is the role of oxygen?
Using a ruthenium on silica-based catalyst (5 wt% Ru) prepared
from a classical incipient wetness impregnation method, a con-
sumption of 18O2 was observed (Fig. 6). No other isotopomer of
oxygen were observed during the reaction showing the absence
of exchange phenomenon. The total pressure in the mass spec-
trometer being maintained constant, the disappearance of 18O2
only involved an increase of the N2 pressure. Then, the resulting
products of these two reactions were analyzed by GC–MS.
Fig. 7 shows the GC–MS analysis of the epoxidation reaction
performed with the POW catalyst. As it was shown in Fig. 1, dur-
ing this epoxidation reaction, no 18O2 consumption is observed.
The GC–MS analysis of the product shows that the MO epoxide is
obtained with two characteristic radical ions at m/z = 155 and at
m/z = 199 corresponding to a 16O epoxide (Fig. 7).
The epoxidation reaction performed with Ruthenium catalyst
was also analyzed by GC–MS (Fig. 8). During this reaction, a con-
sumption of labeled oxygen isotope 18 was observed as previously
shown (Fig. 6). Fig. 8 shows that the GC–MS analysis corresponds
to the MO epoxide is obtained but all the characteristic radical ions
have the value of the standard MO epoxide + 2. In fact the molecu-
Fig. 8. GC–MS analysis of the MO epoxide obtained with the ruthenium based cat-
alyst under a 18O2 flow.

E. Poli et al. / Catalysis Today 157 (2010) 371–377
377
The most probable answer is that O2 initiate the rad-
Acknowledgment
icalar decomposition of H2O2 leading to the formation of
MO mono and diols; this additional parallel reaction path-
way leading to an acceleration of the all epoxidation pro-
cess.
EP thanks the French ministry of education and research for her
PhD grant.
The reuse of the catalyst is still a problem while its separa-
tion from the reaction media is not possible without degradation.
To solve this problem, we studied the heterogeneisation of
the active species; which will be the object of future publica-
tions.
References
[1] J.F. Jenck, F. Agterberg, M.J. Droescher, Green Chem. 6 (2004) 544.
[2] A. Corma, S. Iborra, A. Velty, Chem. Rev. 107 (2007) 2411.
[3] B.K. Sharma, K.M. Doll, S.Z. Erhan, Green Chem. 9 (2007) 469.
[4] U. Biermann, W. Friedt, S. Lang, W. Lubs, G. Machmuller, J.O. Metzger, M.R.
Klaas, H.J. Schafer, M.P. Schneider, Angew. Chem. Int. Ed. Engl. 39 (2000) 2206.
[5] B.S. Lane, K. Burgess, Chem. Rev. 103 (2003) 2457.
[6] G.J.H. Grigoropoulou, J.A. Elings, Green Chem. 5 (2003) 1 (and references
therein).
4. Conclusion
[7] M.A. Camblor, A. Corma, P. Esteve, A. Martinez, S. Valencia, Chem. Commun. 8
(1997) 795.
[8] B. Notari, Adv. Catal. 41 (1996) 253.
[9] D.E. de Vos, B.F. Sels, M. Reynaers, Y.V. Subba Rao, P. Jacobs, Tetrahedron Lett.
39 (1998) 3221.
[10] K. Sato, M. Aoki, M. Ogawa, T. Hashimoto, D. Panyella, R. Noyori, Bull. Chem.
Soc. Jpn. 70 (1997) 905.
[11] W.A. Herrmann, R.M. Kratzer, H. Ding, W.R. Thiel, H. Glas, J. Organomet. Chem.
555 (1998) 293.
[12] C. Venturello, R. D’Aloisio, J. Org. Chem. 53 (1988) 1553.
[13] H. Baumann, M. Buhler, H. Fochem, F. Hirsinger, H. Zoebelein, J. Falbe, Angew.
Chem. 100 (1998) 41.
We studied the influence of an oxygen flow on the reactivity of
a peroxophosphotungstate catalyst for the epoxidation of methyl
oleate (MO) with hydrogen peroxide. Thanks to labeled oxygen
experiments, we proved that the oxygen atoms coming from the
gas phase were not directly involved in the epoxide formation. We
also detected some new intermediates during this reaction; the
mono and the diol of MO are formed at the early stage of the reac-
tion by a radicalar mechanism. This mechanism is the minor way
of a parallel epoxidation mechanism which also involve a major
concerted mechanism way. The reaction is faster under O2 or air
flow because the radicalar mechanism does not occur under a nitro-
gen flow, and the influence of O2 on the reaction kinetic is due to
its participation in the radicalar transformation of methyl oleate
into its epoxide by way of the formation of the MO mono and diol
intermediates.
[14] E. Santacesaria, A. Sorrentino, F. Rainone, M. Di Serio, F. Speranza, Ind. Eng.
Chem. Res. 39 (2000) 2766.
[15] E. Poli, J.-M. Clacens, J. Barrault, Y. Pouilloux, Catal. Today 140 (2009) 19.
[16] R. Noyori, M. Aoki, K. Sato, Chem. Commun. 16 (2003) 1977.
[17] D. Martin, D. Duprez, J. Chem. Phys. 100 (1996) 9429.
[18] S. Casciato, PhD thesis, Nouveaux catalyseurs solides pour la réaction
d’époxydation d’esters gras insaturés sans solvant, Poitiers, 2009.