A highly efficient, green and recoverable catalytic system for the epoxidation of fatty esters and biodiesel with H2O2

Article abstract of DOI:10.1016/j.apcata.2012.03.009

A highly efficient and green strategy for the epoxidation of fatty esters, combining a green oxidant (aqueous hydrogen peroxide) and a recyclable catalyst (a seleninic acid derivative) is presented. The possibility of integrating renewable solvents derived from glycerol in the productive cycle of biodiesel commodities is also explored. Fluorinated solvents (both commercial and glycerol-derived) play a double key role on this methodology: on the one hand, they strongly accelerate epoxidation reaction with respect to common non-fluorinated solvents and, on the other hand, some of them allow catalyst recycling. Mono- and polyunsaturated fatty esters, as well as soy biodiesel have been epoxidized with excellent yields and selectivities towards epoxide under the optimal conditions and with minimum catalyst loads.

Full text of DOI:10.1016/j.apcata.2012.03.009

Applied Catalysis A: General 425–426 (2012) 91–96
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
A highly efficient, green and recoverable catalytic system for the epoxidation of
fatty esters and biodiesel with H2O2
Miriam De Torresa,1, Isabel W.C.E. Arendsb,2, José A. Mayoralc,3, Elisabet Piresa,1
,
Gonzalo Jiménez-Osésa,d,∗
a Departamento de Química Orgánica, Instituto de Ciencia de Materiales de Aragón, Facultad de Ciencias, Universidad de Zaragoza-CSIC, Pedro Cerbuna, 12, E-50009 Zaragoza, Spain
b Laboratory for Biocatalysis and Organic Chemistry, Delft University of Technology, Julianalaan 13, 2628 BL Delft, Netherlands
c Instituto Universitario de Catálisis Homogénea, Facultad de Ciencias, Universidad de Zaragoza, Pedro Cerbuna, 12, E-50009 Zaragoza, Spain
d Department of Chemistry and Biochemistry, University of California, Los Angeles, California, 607 Charles E. Young Drive East, 90095, Los Angeles, CA, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
A highly efficient and green strategy for the epoxidation of fatty esters, combining a green oxidant (aque-
ous hydrogen peroxide) and a recyclable catalyst (a seleninic acid derivative) is presented. The possibility
of integrating renewable solvents derived from glycerol in the productive cycle of biodiesel commodities
is also explored. Fluorinated solvents (both commercial and glycerol-derived) play a double key role on
this methodology: on the one hand, they strongly accelerate epoxidation reaction with respect to com-
mon non-fluorinated solvents and, on the other hand, some of them allow catalyst recycling. Mono- and
polyunsaturated fatty esters, as well as soy biodiesel have been epoxidized with excellent yields and
selectivities towards epoxide under the optimal conditions and with minimum catalyst loads.
© 2012 Elsevier B.V. All rights reserved.
Received 5 January 2012
Received in revised form 2 March 2012
Accepted 3 March 2012
Available online 12 March 2012
Keywords:
Epoxidation
Biodiesel
Hydrogen peroxide
Selenium
Fluoroalcohols
1. Introduction
selectivities, because of the poorer reactivity of H2O2 compared to
other classic epoxidizing reagents such as organic peracids, perox-
In the last decade great interest has been focused in the devel-
opment of new efficient processes for the functionalization of
fatty acids through oxidation reactions [1]. This interest has been
prompted by the need of reaching renewable chemicals in order
to satisfy society demands for a greener chemistry. In this way,
epoxidation of fatty acids derivatives is one of the most promising
reactions to achieve the synthesis of a large variety of interesting
products such as hydroxyethers, hydroxyesters, diols, etc. To this
aim, the use of hydrogen peroxide (H2O2) for selective epoxida-
tions is highly desirable because it is cheap, it has a high content
in active oxygen, and it is clean, since the only byproduct formed
is water [2]. The use of H2O2 as oxidant often requires an effi-
cient catalytic process in order to achieve high conversions and
yesters or persulfates. Many catalytic systems based on different
metals have been reported for the epoxidation of a wide range of
alkenes using hydrogen peroxide [3].
With respect to fatty acid derivatives epoxidation, several exam-
ples using the classical conditions based on peroxy acids generated
in situ from carboxylic acids (typically formic or acetic acids) as
active oxygen carriers, H2O2 as oxygen donor and strong mineral
acids (H2SO4, HNO3, H3PO3) as catalysts have been reported [4].
Inorganic and organometallic catalysts based on tungsten [5], rhe-
nium and manganese [6] have also been used. The heterogeneous
version of the epoxidation reaction with H2O2 on fatty derivatives
has also been explored using amorphous SiO2/Ti [7] and methyl-
trioxorhenium on niobia catalysts [8], ion exchange resins [9], and
sol–gel aluminas [10]. In the past years, the enzimatically promoted
epoxidation of vegetable oils and fatty acids is gaining increasing
interest [11].
Aryl diselenides [12] and their oxidized counterparts, i.e. aryl
seleninic [13] and peroxyseleninic acids [14] have been described
as highly efficient oxygen transfer catalysts for different processes,
principally the epoxidation of olefins. Heterogenized versions of
these catalysts have been also described [15]. On the other hand, the
so-called “booster effect” of fluorinated alcohols, which has been
beautifully described by the groups of Neumann and co-workers
∗
Corresponding author at: Department of Chemistry and Biochemistry, Univer-
sity of California, Los Angeles, California, 607 Charles E. Young Drive East, 90095, Los
Angeles, CA, United States. Tel.: +1 310 2060515; fax: +1 310 2061843.
E-mail addresses: I.W.C.E.Arends@tudelft.nl (I.W.C.E. Arends),
mayoral@unizar.es (J.A. Mayoral), epires@unizar.es (E. Pires),
gjimenez@chem.ucla.edu (G. Jiménez-Osés).
1
Tel.: +34 976762271; fax: +34 976762077.
Tel.: +31 15 2784423; fax: +31 15 2784700.
Tel.: +34 976762077; fax: +34 976762077.
2
3
0926-860X/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2012.03.009

92
M. De Torres et al. / Applied Catalysis A: General 425–426 (2012) 91–96
[16], van Vliet et al. [17] and Berkessel et al. [18], allows the use
of H2O2 in the absence of a catalyst to epoxidize different types of
alkenes, although usually at the expense of longer reaction times
and higher temperatures. These studies have demonstrated the
activation of H2O2 in fluorinated solvents towards oxidation reac-
tions increasing reaction rates up to five times.
The combination of both aryl diselenides as catalysts and fluo-
rinated solvents has raised a powerful methodology, developed by
Sheldon and co-workers, for the epoxidation of cyclic and acyclic
alkenes and the oxidation of carbonyl compounds, with very high
turnover numbers and selectivities under different mono-, bi-, and
even triphasic conditions [19]. Our aim was to test the applicabil-
ity of this methodology, and eventually to improve it, to efficiently
epoxidize fatty esters and biodiesel, in competition with the exist-
ing common methodologies.
In this sense, an important shortcoming of many of these epoxi-
dation methods is, however, that catalysts are not readily available
and that some of them are expensive to prepare and potentially
toxic. Thus, greenness and economy of the catalytic reactions can
be further improved by catalyst recycling, which is usually accom-
plished by using solid heterogeneous catalysts. However, the use of
classical heterogeneous catalysts (for instance, based on titanium)
together with aqueous hydrogen peroxide often results in exten-
sive epoxide hydrolysis, and many efforts have been devoted to the
synthesis of heterogeneous solid catalysts with hydrophobic prop-
erties [20]. A possible alternative would be the use of homogeneous
catalysis in a liquid phase allowing easy separation of reagents and
products by distillation from the catalytic phase, which would be
subsequently recycled for further reactions. This strategy has been
described for other reactions in the case of ionic liquids [21]. In a
previous work, we described the homogeneously catalyzed epox-
idation reaction of cyclooctene, using aqueous hydrogen peroxide
as oxidant and a diphenyl diselenide as catalyst precursor in a
wide range of solvents [22]. Although classically this catalytic sys-
tem had been dominated by the use of dichloromethane (DCM)
and 2,2,2-trifluoroethanol (TFE) [12b,19], it was found that some
glycerol-derived solvents are as good as TFE in some cases and bet-
ter than DCM in others, the best ones being those containing short
fluorinated chains. These glycerol-derived solvents, which present
some attractive advantages, such as being renewable, presumably
low toxic, low volatile, and possessing tunable physico-chemical
properties [23], have ultimately allowed the recycling of the cat-
alytic phase thorough the selective evaporation of volatile reaction
products, i.e. cyclohexene oxide [24]. However, one important
drawback of this attractive strategy is being limited to volatile
reagents and products, which is not the case of unsaturated fatty
derivatives as oleates and linoleates.
Scheme 1. Catalytic epoxidation of unsaturated fatty methyl esters with arylse-
leninic acid–hydrogen peroxide system in several common and glycerol-derived
solvents (fluorinated and no-fluorinated).
Then, hydrogen peroxide (60% in water, 2 mmol, 0.1 mL) was
added and the mixture was heated at reflux or 90 ◦C (in high-
boiling point GlyF). Reactions were monitored by taking samples
at regular times. Those samples were dissolved in tetrahydrofu-
ran; some manganese dioxide was added to quench the excess of
hydrogen peroxide in the sample. Samples were dried over anhy-
drous MgSO4 and analyzed by GC. Glycerol-based solvents were
obtained by ring-opening of epichlorohydrin with the correspond-
ing alkoxide (sodium ethoxide) or alcohol (TFE) in alcoholic media
[23] and subsequent purification by vacuum distillation.
2.2. Se-catalyzed reactions
The general procedure for the epoxidation of fatty acid methyl
esters was carried out at 25 ◦C, using the following conditions:
0.1–1.0 mol% per mol4 of unsaturation of bis[3,5-bis(trifluoro-
methyl)diphenyl] diselenide [14a,e], was dissolved in 1 mL of
solvent. Next, hydrogen peroxide (60% in water, 2 mol per mole
of unsaturation) was added. When the solution became colorless,
eicosane (internal standard, 0.15 mmol, 40 mg) was added followed
by 1 mmol of the olefin. The reaction was monitored by GC as
described above.
For biodiesel epoxidation, the nature and proportion of unsatu-
rated methyl esters was considered (typically, 50% of methyl oleate,
30% of methyl linoleic and 5% of methyl linolenate in soy biodiesel).
Once the total unsaturation grade is calculated (1.45 mmol/300 mg
in our biodiesel samples), the epoxidation reaction is carried out
under the same conditions described above.
In this paper, we overcome this limitation and take advan-
tage of using
a green oxidant (hydrogen peroxide) and a
highly active homogeneous catalyst (namely bis[3,5-bis(trifluoro-
methyl)diphenyl] diselenide) [14a,e] in combination with fluori-
nated alcohols as solvents or co-solvents, to selectively epoxidize
fatty methyl esters and their mixtures as presented in biodiesel
(Scheme 1). The use of adequate amounts of fluorinated alcohols is
of key importance to both accelerate the reaction and to prepare
recyclable catalytic liquid phases.
2.3. Recycling experiments
Bis[3,5-bis(trifluoro-methyl)diphenyl] diselenide (0.01 mol,
5.8 mg) was dissolved in 5 mL of TFE. Next, hydrogen peroxide
(60% in water, 1,1 mmol, 0.06 mL per cycle) was added. When the
solution became colorless, eicosane (internal standard, 0.15 mmol,
40 mg per cycle) was added followed by methyl oleate (1 mmol,
297 mg per cycle). The reaction was monitored by GC as described
2. Experimental
2.1. Non-catalyzed reactions
Methyl oleate (1 mmol, 297 mg) and eicosane (internal standard,
0.15 mmol, 40 mg) were dissolved in 1 mL of the corresponding
fluorinated alcohol.
4
Although the optimum catalyst load is 0.1 mol% when HFIP is used as a sol-
vent, the use of 1 mmol% is strongly recommended in order to maintain reasonable
reaction times along the successive reaction cycles.

M. De Torres et al. / Applied Catalysis A: General 425–426 (2012) 91–96
93
above. When the reaction was complete, the products and internal
standard were extracted with isooctane (3 × 3 mL, upper layer),
dried over anhydrous MgSO4, filtered and evaporated under
reduced pressure. Then more hydrogen peroxide and eicosane
were added to the remaining residue (solvent and catalyst). After
a few minutes more olefin (1 mmol, 297 mg) was added. When the
last reaction cycle is complete, the catalyst is completely removed
˚
by entrapment in 4 A molecular sieves powder and filtration.
3. Results and discussion
3.1. Non-catalyzed epoxidation
Given the remarkable effects of fluorinated alcohols, as men-
tioned in the Introduction, and inspired on previous works in
which TFE, 1,1,1,3,3,3-hexafluoroisopropanol (HFIP) [16–18] and
1,3-bis(2,2,2-trifluoroethoxy)-2-propanol (GlyF) [22] were suc-
cessfully used as activating solvents in the absence of catalysts for
the epoxidation of diverse olefins with H2O2, we first assayed the
direct epoxidation of methyl oleate with H2O2. An initial screen-
ing of common alkyl and fluoroalkyl alcohols was carried out to
determine the most suitable solvent in terms of rate, conversion
of alkene and selectivity towards the corresponding oxide. Apart
from the commercially available solvents, glycerol-derived sym-
metric ethers such us 1,3-diethoxypropan-2-ol and GlyF were first
prepared and purified following a protocol previously reported by
our group [23]. DCM was selected as a non-alcoholic probe, which
in turn is widely used in the environmentally harmful epoxidation
of alkenes with organic perbenzoic acids like m-chloroperbenzoic
acid (MCPBA). In this initial survey, an excess of H2O2 (60% in
water) was employed (2 equiv). Under these conditions, the reac-
tion at room temperature did not progress at all or was too slow
for being analyzed, so the reaction temperature was raised up to
reflux conditions for all solvents and up to 90 ◦C for GlyF. When
Fig. 1. Solvent effects in the Se-catalyzed epoxidation of methyl oleate. Reaction
conditions: 1 mmol methyl oleate, 2 eq. H2O2, 1 mol% catalyst 1, 1 mL solvent, room
temperature. Yields determined by GC.
temperature-dependent products distribution were observed. The
corresponding allyl alcohols, 1,2-diol and 1,2-hidroxyethers result-
ing from allylic oxidation and the ring-opening reactions of the
oxirane with water and GlyF, respectively, were identified among
these products. Apart from these expected side reactions, which
have been observed before under similar conditions [16a,22], the
rather complex mixture of products obtained could be related to
the thermal decomposition of H2O2, which seems to quickly take
place at temperatures above 90 ◦C (i.e., no reaction occurred at
80 ◦C). Evidence pointing to a radical-mediated epoxidation result-
ing from the thermal homolytic cleavage of hydrogen peroxide is
that variable amounts of cis and trans epoxides were obtained [25].
Noteworthy, a major proportion of the trans epoxide5 was obtained
even at low conversion values (40% conv at 22 h, 2:1 trans/cis ratio)
when the non-catalyzed epoxidation reaction was conducted under
the same conditions but using the non-fluorinated analogue of GlyF
(1,3-diethoxypropan-2-ol) as a solvent. In this case, much lower
conversions and a similar complex mixture were obtained. These
results, together with those described for the Se-catalyzed reaction
(vide infra), clearly discards this compound as a suitable solvent
to perform the epoxidation of unsaturated fatty esters, despite is
renewable character.
i
commercial non-fluorinated alcohols (EtOH, PrOH, tBuOH) were
used as solvents in the absence of Se catalyst (1), no detectable
amount of epoxide was obtained after prolonged heating, which
revealed their much worse properties as hydrogen peroxide activa-
tors. Conversely, HFIP was revealed as the most convenient solvent
for this process, cleanly leading to 76% yield of cis-epoxide in 2 h
and to a full completion of the reaction in 5 h. TFE exhibited slower
reaction rates, yielding only 24% yield in 4.5 h and being neces-
sary 28 h of heating at reflux to reach 71% yield of epoxide in the
same cis configuration. It must be noted that the reflux temper-
ature of HFIP (59 ◦C) is lower than that of TFE (77–80 ◦C), so the
outstanding accelerating properties of the former are corroborated.
The crude mixtures obtained using these commercial fluorinated
alcohols as solvents were very clean under the standard conditions
(1 mmol of methyl oleate, 2 mmol of H2O2 and 1 mL of solvent).
In an attempt to reduce reaction times with TFE, an excess of 4
equiv. of H2O2 was used, which slightly increased epoxide yield at
short reaction times (38% versus 24% after 4.5 h) but, in turn, did
not allowed to reach full conversion of methyl oleate. Worse still,
significant amount of the corresponding 1,2-diol (43%) and traces
of the 1,2-(trifluoroethoxy)alcohol were generated at the expense
of reducing the global yield of epoxide (30%) through side ring-
opening reactions, after 20 h of stirring under reflux (80% conv).
These results, together with those obtained in the recycling assays
(vide infra) show that long reaction times, more than a large excess
of water in the reaction mixture, seriously affect the selectivity of
the epoxidation reaction and should be, as far as possible, avoided.
For this reason, the use of GlyF under non-catalytic conditions
was revealed to be problematic, due to the poor selectivity achieved
towards epoxidation. With this solvent, nearly all the starting mate-
rial is consumed after prolonged heating at reflux, but time and
3.2. Se-catalyzed epoxidation
After demonstrating the different accelerating properties of the
solvents in the control reactions described above, we thoroughly
assayed the selenium-catalyzed epoxidation of fatty acid deriva-
tives, the main objective of our study. Initially, the amount of
catalyst precursor 1 was fixed to 1 mol% (i.e., 2 mol% of the cat-
alytically active species, Scheme 1), and the excess of H2O2 (60%
in water) employed was again fixed to 2 equiv. The same solvents
used in the non-catalytic tests were employed. Reactions were car-
ried out at room temperature and under concentrated conditions,
namely 1 mL of solvent and 1 mmol of substrate (methyl oleate). As
could be confirmed by GC and 1H NMR analysis of the crude mix-
tures, a complete selectivity towards the corresponding cis epoxide
was obtained in all cases. The results of this study are summarized
in Fig. 1. As can be seen on these plots, and according to previ-
ous reports [16–18] and the control results described above, the
5
The epoxide in trans configuration was selectively obtained, characterized in
parallel by applying the same Se-catalyzed oxidation protocol using methyl elaidate
((E)-methyl octadec-9-enoate as starting material.

94
M. De Torres et al. / Applied Catalysis A: General 425–426 (2012) 91–96
Table 1
epoxidation reaction was greatly accelerated in the presence of
fluorinated alcohols, being the maximum final turnover frequen-
cies (TOFF, measured when the substrate was totally consumed)
achieved with HFIP (TOFF = 588 h−1) and TFE (TOFF = 67 h−1). It must
be emphasized that this acceleration cannot be attributed to a
significant contribution of the non-catalyzed epoxidation reaction
pathway, given the much slower reaction rates described in the
precedent section, even at higher reaction temperatures. Accelera-
tion by solvent must be therefore connected to the catalyst action.
Noteworthy, using a fluorinated glycerol-derived solvent such
as 1,3-bis(2,2,2-trifluoroethoxy)-2-propanol (GlyF, TOFF = 16 h−1),
despite bearing rather separated fluorine and hydroxyl groups,
allows to perform the selective epoxidation of fatty esters with
excellent yields in a short time and with limited wastes. This strat-
egy allows to efficiently integrate all the materials derived from
the production of biodiesel, that is, the fatty methyl esters and glyc-
erol, into the same process, which would contribute to improve the
productive cycle of a biorefinery. With respect to non-fluorinated
solvents, only DCM allows an acceptable reactivity (TOF = 12 h−1),
in good agreement with previous reports [22]. On the other side,
all non-fluorinated alcohols showed low conversions at short times,
being necessary at least 24 h to reach only a 75% yield under the best
conditions (iPrOH). Thus, glycerol-derived 1,3-diethoxypropan-2-
ol, albeit being more environmentally and economically convenient
than its fluorinated analogue (GlyF), is not suitable to perform the
epoxidation of unsaturated fatty compounds, in view of the slow
reaction rates and low yields obtained.
At this early stage of the study it was also noticed that, accord-
ing to recent observations [26], the time required to preactivate
the catalyst, i.e. to oxidize the diselenide (pale yellow) to the
seleninic acid (colorless) in the presence of H2O2, correlated with
the global reaction time (from 15 min of preactivation in iPrOH to
5 min in TFE to 20 s in HFIP). The advantage of fluoroalcohols in
similar Se-catalyzed reactions has been attributed to their polar
but non-basic, non-oxidisable and water-miscible character, which
avoids mass transfer limitations and reduces epoxide solvolysis
rate [12b]. Furthermore, the results described herein demonstrate
that the accelerating properties of fluoroalcohols also play a key
role in the H2O2-promoted activation (and regeneration) of the cat-
alyst. However, this activation was not strong enough to promote
the regeneration of MCPBA with H2O2 (which commonly require
strongly acidic catalysts and/or solvents [27]) in TFE as solvent
under the same conditions, precluding the use of catalytic amounts
MCPBA in a similar manner.
Optimization of the reaction conditions for the epoxidation of methyl oleate with
H2O2 using diselenide 1 as catalyst at room temperature.
Entry
Solvent
H2O2 (equiv.)
Catalyst (mol%)
Yielda
Time
1
2
3
4
5
6
7
8
TFE
TFE
2.0
2.0
2.0
2.0
2.0
2.0
1.1
1.1
1.0
0.5
1.0
0.5
0.25
0.1
1.0
1.0
>99%
>99%
>99%
>99%
>99%
>99%
87%
1.5 h
2 h
HFIP
HFIP
HFIP
HFIP
TFE
10 min
15 min
15 min
20 min
1.5 h
HFIP
93%
10 min
a
Determined by GC.
0.5% slightly increased the total reaction time in TFE from 1.5 to 2 h
(Table 1, entries 1 and 2). More impressive was the significant ten-
fold reduction of the catalyst load that could be accomplished when
carrying out the reaction in HFIP, in which with only 0.1 mol% of 1
the epoxidation of methyl oleate was completed in 25 min (Table 1,
entries 3–6).
Despite its outstanding accelerating capacity, the high price of
HFIP can be a serious drawback to perform the reaction in a cost-
effective manner at an industrial scale. For this reason, we assayed
the use of decreasing volumes of HFIP as a cosolvent in the presence
of its non-fluorinated and inexpensive analogue (iPrOH). Catalyst
1 (1 mol%) and 1 mL of total volume with variable proportions of
iPrOH and HFIP were used. As can be seen (Table 2, entries 1–3),
a reasonable catalytic activity can be maintained with 50% (v/v)
of HFIP, although longer reaction times are required to complete
the reaction. The same trend is observed with glycerol-derived 1,3-
diethoxypropan-2-ol, albeit its flourinated analogye (GlyF) shows a
superior performance (Table 2, entries 4 and 5), in good agreement
with the results described above. Presumably, hydroxylated alkyl
cosolvents could interfere within the hydrogen bonding network
created by HFIP around hydrogen peroxide [18], so we decided to
repeat these experiments but with other widely used non-alcoholic
solvents such us n-hexane, toluene, ethyl acetate, tetrahydrofuran
and tert-butyl methylether. (Table2, entries6–16) Accordingto this
hypothesis, a good correlation between hydrogen donor/acceptor
capabilities of each solvent/cosolvent and their accelerating prop-
erties was observed. Non-polar hydrocarbon solvents (n-hexane,
toluene) were the best co-solvents in terms of reactivity, followed
The scope of this non-optimized selenium-catalyzed epoxi-
dation protocol was quickly evaluated by performing the same
reaction using methyl linoleate and soybean biodiesel as substrates
and industrially affordable TFE. The amount of hydrogen peroxide
was carefully adjusted depending on the number of unsaturations
and the nature and proportion of the unsaturated compounds (see
Section 2). These experiments revealed that the epoxidation of
methyl oleate and the first double bond of poly-unsaturated fatty
esters were rather fast because of the greater concentration of
hydrogen peroxide at short reaction times (∼30 min). On contrary,
the complete epoxidation of the second double bond of methyl
linoleate took place much more slowly (2–2.5 h), when at least one
equiv. of H2O2 had been consumed, and the formation of related
ring-opening diols was observed at long reaction times (3.5–4 h).
Therefore, it was shown that this selenium/fluoroalcohol-based
methodology can be successfully applied for the epoxidation of
both pure and mixtures of mono- and polyunsaturated compounds
under very mild and sustainable conditions.
Table 2
Solvent mixtures for the epoxidation of methyl oleate with H2O2 using diselenide 1
as catalyst.a
Entry
Solvent
Yieldb (time)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
HFIP/iPrOH (1:1)
61% (30 min), 97% (3 h)
55% (3 h), >99% (23 h)
39% (3 h), 88% (23 h)
49% (30 min)
91% (30 min)
97% (15 min)
95% (15 min)
98% (1.5 h)
97% (15 min)
94% (15 min)
57% (15 min), 92% (1 h)
56% (30 min), 79% (1 h)
78% (5 h)
35% (3 h), 83% (23 h)
27% (30 min)
HFIP/iPrOH (1:3)
HFIP/iPrOH (1:9)
HFIP/1,3-diethoxypropan-2-ol (1:1)
HFIP/GlyF (1:1)
HFIP/n-hexane (1:1)
HFIP/n-hexane (1:3)
HFIP/n-hexane (1:9)
HFIP/toluene (1:1)
HFIP/toluene (1:3)
HFIP/toluene (1:9)
HFIP/ethyl acetate (1:1)
HFIP/ethyl acetate (1:3)
HFIP/ethyl acetate (1:9)
HFIP/THF (1:1)
HFIP/tBuOMe (1:1)
48% (30 min)
Once the best solvents to perform de epoxidation of methyl
oleate were found (TFE and HFIP), an optimization of the catalyst
load was done in order to improve the economy of the process.
In this sense, the reduction in the amount of catalyst from 1% to
a
Reaction conditions: 1 mmol methyl oleate, 2 equiv. H2O2, 1 mol% catalyst 1,
1 mL solvent, room temperature.
b
Determined by GC.

M. De Torres et al. / Applied Catalysis A: General 425–426 (2012) 91–96
95
has been reported on previous studies [26]. This problem could
be partially overcome by diluting methyl oleate fivefold in TFE, at
an expense of a more prolonged reaction time (2 h) to reach full
conversion. The increase of reaction volume (5 mL) also facilitated
the extractive workup, so that this operation could be optimized
to achieve a global >99% yield of isolated epoxide after 3 × 3 mL
extractions with isooctane. As represented in Fig. 2 (Method B),
the same catalytic phase could be reused up to five times although
an increasing loss of activity due to the accumulation of water is
observed. For this reason, the amount of H2O2 was first reduced
from 2 eq. to 1.1 eq., as described before (Table 1, entries 7 and
8). Additionally, the common strategy of adding a strong drying
˚
agent like 4 A molecular sieves (200 mg per mmol of substrate) to
remove water from the reaction phase after the extraction of prod-
ucts with isooctane, resulted in a complete loss of activity after
filtering and washing the solid with TFE.7 This issue was due to the
complete entrapment of the catalyst in the solid, as demonstrated
by Selenium analysis of the recovered molecular sieves. On con-
trary, the same treatment using optimized amounts of anhydrous
MgSO4 after the fourth reaction cycle (400 mg per mmol of sub-
strate) allowed to restore the initial catalytic activity (Fig. 2, Method
C). With this optimized recovering protocol, the catalytic phase
could be reused up to seven times without significant loss of activ-
ity or cis stereoselectivity. A final treatment of the fluorous phase
with molecular sieves allows the complete removal of the catalyst
for its safe disposal. In summary, an important conclusion drawn
from these recoverability assays, is that water does not deactivate
Se/fluoroalcohols catalytic systems irreversibly; instead, an indi-
rect and reversible decrease on the reaction rate due to solubility
and mass transfer issues is produced.
Fig. 2. Catalyst recovery experiments. Method A: 1 mmol methyl oleate, 2 equiv.
H2O2, 1 mol% catalyst 1, 1 mL TFE, room temperature. Method B: 1 mmol methyl
oleate, 2 equiv. H2O2, 1 mol% catalyst 1, 5 mL TFE, room temperature. Method C:
1 mmol methyl oleate, 2 equiv. H2O2, 1 mol% catalyst 1, 5 mL TFE, room temperature.
Extractions were performed with isooctane (3 × 3 mL). Yields determined on the
crude mixture after extraction and evaporation.
by hydrogen bond acceptors (ethers6 and esters) and hydrogen
bond donors/acceptors (alcohols).
As found thorough these studies, excellent epoxide yields in
relatively short times and in the presence of non-toxic and inexpen-
sive solvents, together with small amounts of HFIP, can be obtained.
It must be noted that, with 10% (v/v) HFIP solvent mixtures,
only 0.95 mmol of this fluoroalcohol per mole of substrate is
required to notably accelerate the epoxidation reaction, which is
an excellent compromise between efficiency and economy. These
unprecedented results reveal a true catalytic role of such fluori-
nated solvents, and expand the scope of these derivatives to be
used in industry in a sustainable manner.
4. Conclusions
As a final step, the excess of hydrogen peroxide was optimized.
Excellent yields of epoxide (87–93%) with total cis-selectivity were
obtained using 1.1 equiv of H2O2 in TFE and HFIP, respectively
under the same conditions (Table 1, entries 7–8).
This allowed to significantly reduce the amount of water gen-
erated as byproduct, which was of capital importance to achieve a
good catalyst recoverability, as will be described later on.
A highly efficient and green strategy for the epoxidation of fatty
esters, combining a green oxidant (aqueous hydrogen peroxide)
and a recyclable catalyst (a seleninic acid derivative) has been
explored. The possibility of integrating renewable solvents derived
from glycerol in the productive cycle of biodiesel commodities is
also tested. Fluorinated solvents (both commercial and glycerol-
derived) play a key double role on this methodology: on the one
hand, they strongly accelerate epoxidation reaction with respect
to common non-fluorinated solvents and, on the other hand, some
of them allow the easy catalyst recycling by extraction of the reac-
tion products with inexpensive alkane solvents like hexanes and
isooctane, and subsequent reuse of the fluorinated catalytic phase.
The activating properties of certain fluorinated solvents (TFE and
especially HFIP) towards hydrogen peroxide are so strong that the
epoxidation of fatty derivatives can be accomplished even in the
absence of a catalyst, albeit at the expense of a longer reaction times
and higher temperatures. A negative effect of water accumulation
has been observed, although it does not result in irreversible cata-
lyst deactivation and can be overcome by the use of an inexpensive
desiccant. Great optimization efforts have been done to achieve
an environmentally friendly and sustainable protocol. In summary,
mono- and polyunsaturated fatty, as well as soy biodiesel have been
epoxidized with excellent yields and selectivities towards epox-
ide under the optimal conditions with minimum catalyst loads
(0.1–1% mmol) and the possibility of reusing the catalytic system
up to seven reaction cycles.
3.3. Catalyst recycling experiments
Once the best reaction conditions for the epoxidation of fatty
methyl esters were found, we focused our attention on recycling
and/or reusing the catalyst. To this aim, we found that some com-
mon and inexpensive alkanes were practically immiscible with
the reaction mixture in TFE. This allowed to take advantage of
the highly hydrophobic character of the fatty reaction products, in
order to isolate them by extraction with these alkanes. As demon-
strated by preliminary assays, isooctane (2,2,4-trimethylpentane)
showed superior extracting properties than hexanes and was the
solvent of choice for this process, in which the catalyst remains
‘inmobilized’ in the fluoroalcoholic phase, ready to be used for addi-
tional reaction cycles after recharging with more substrate and
oxidant. However, it was found that the standard protocol based
on substrate concentrations of 1 M in TFE could not be repeated
more than four times due to the accumulation of water on the cat-
alytic phase and the appearance of insolubility and mass transfer
problems associated to it (Fig. 2, Method A). The negative effect of
the increasing amount of water formed along the oxidation cycles
7
While activated molecular sieves powder fully disrupted the catalytic activity
6
As solvent grade tetrahydrofuran was used, the presence of water traces could
of 1, beads did not show any beneficial effect and the reaction proceeded as in the
be the cause of the observed poor performance.
absence of the desiccant.

96
M. De Torres et al. / Applied Catalysis A: General 425–426 (2012) 91–96
Acknowledgments
(b) A.G. Kutateladze, J.L. Kice, T.G. Kutateladze, N.S. Zefirov, J. Org. Chem. 58
(1993) 995–996;
(c) L.G. Faehl, J.L. Kice, J. Org. Chem. 44 (1979) 2357–2361;
(d) T. Hori, K.B. Sharpless, J. Org. Chem. 43 (1978) 1689–1697;
(e) J.L. Kice, T.W.S. Lee, J. Am. Ceram. Soc. 100 (1978) 5094–5102.
[14] (a) L. Syper, J. Mlochowski, Tetrahedron 43 (1987) 207–213;
(b) H.J. Reich, F. Chow, S.L. Peake, Synthesis (1978) 299–301;
(c) L. Syper, Synthesis (1989) 167–172;
Financial support from the European Union (project Solvsafe,
NMP2-CT-2005-011774) and the Spanish MCIMM (projects
CTQ2008-05138 and Consolider Ingenio 2010 CSD 2006-0003) is
gratefully acknowledged. M. de Torres thanks the Gobierno de
Aragón for a grant. G. Jiménez-Osés acknowledges Gobierno de la
Rioja for a Postoctoral Fellowship.
(d) L. Syper, Tetrahedron 43 (1987) 2853–2871;
(e) L. Syper, J. Mlochowski, Synthesis (1984) 747–751.
[15] (a) R.T. Taylor, L.A. Flood, J. Org. Chem. 48 (1983) 5160–5164;
(b) B. Betzemeier, F. Lhermitte, P. Knochel, Synlett (1999) 489–491;
(c) H. Wojtowicz, G. Soroko, J. Mlochowski, Synth. Commun. 38 (2008)
2000–2010;
References
(d) D. Crich, Y.K. Zou, Org. Lett. 6 (2004) 775–777;
[1] A. Kockritz, A. Martin, Eur. J. Lipid Sci. Technol. 110 (2008) 812–824.
[2] I. Arends, R.A. Sheldon, Top. Catal. 19 (2002) 133–141.
[3] (a) G. Grigoropoulou, J.H. Clark, J.A. Elings, Green Chem. 5 (2003) 1–7;
(b) D.E. De Vos, B.F. Sels, P.A. Jacobs, Adv. Synth. Catal. 345 (2003) 457–473;
(c) D.E. De Vos, B.F. Sels, P.A. Jacobs, Adv. Catal. 46 (2001) 1–87.
[4] (a) V.V. Goud, A.V. Patwardhan, N.C. Pradhan, Bioresour. Technol. 97 (2006)
1365–1371;
(e) D. Crich, Y.K. Zou, J. Org. Chem. 70 (2005) 3309–3311.
[16] (a) K. Neimann, R. Neumann, Org. Lett. 2 (2000) 2861–2863;
(b) S.P. de Visser, J. Kaneti, R. Neumann, S. Shaik, J. Org. Chem. 68 (2003)
2903–2912.
[17] M.C.A. van Vliet, I. Arends, R.A. Sheldon, Synlett (2001) 248–250.
[18] (a) A. Berkessel, J.A. Adrio, Adv. Synth. Catal. 346 (2004) 275–280;
(b) A. Berkessel, J.A. Adrio, J. Am. Ceram. Soc. 128 (2006) 13412–13420;
(c) A. Berkessel, J.A. Adrio, D. Huettenhain, J.M. Neudorfl, J. Am. Ceram. Soc. 128
(2006) 8421–8426.
(b) S. Dinda, A.V. Patwardhan, V.V. Goud, N.C. Pradhan, Bioresour. Technol. 99
(2008) 3737–3744;
(c) V.V. Goud, S. Dinda, A.V. Patwardhan, N.C. Pradhan, Asia-Pac. J. Chem. Eng.
5 (2010) 346–354.
[19] (a) G.J. ten Brink, J.M. Vis, I. Arends, R.A. Sheldon, J. Org. Chem. 66 (2001)
2429–2433;
[5] E. Poli, J.M. Clacens, J. Barrault, Y. Pouilloux, Catal. Today 140 (2009) 19–22.
[6] G.D. Du, A. Tekin, E.G. Hammond, L.K. Woo, J. Am. Oil Chem. Soc. 81 (2004)
477–480.
(b) G.J. ten Brink, J.M. Vis, I. Arends, R.A. Sheldon, Tetrahedron 58 (2002)
3977–3983.
[20] (a) See, for instance T. Blasco, M.A. Camblor, A. Corma, P. Esteve, J.M. Guil, A.
Martinez, J.A. Perdigon-Melon, S. Valencia, J. Phys. Chem. B 102 (1998) 75–88;
(b) J.M. Fraile, J.I. Garcia, J.A. Mayoral, E. Vispe, J. Catal. 189 (2000) 40–51;
(c) J.M. Fraile, J.I. Garcia, J.A. Mayoral, E. Vispe, J. Catal. 204 (2001) 146–156;
(d) J.M. Fraile, J.I. Garcia, J.A. Mayoral, E. Vispe, D.R. Brown, M. Naderi, Chem.
Commun. (2001) 1510–1511;
[7] A. Campanella, M.A. Baltanas, M.C. Capel-Sanchez, J.M. Campos-Martin, J.L.G.
Fierro, Green Chem. 6 (2004) 330–334.
[8] A.O. Bouh, J.H. Espenson, J. Mol. Catal. A: Chem. 200 (2003) 43–47.
[9] (a) R. Mungroo, N.C. Pradhan, V.V. Goud, A.K. Dalai, J. Am. Oil Chem. Soc. 85
(2008) 887–896;
(b) S. Sinadinovic-Fiser, M. Jankovic, Z.S. Petrovic, J. Am. Oil Chem. Soc. 78 (2001)
725–731.
[10] J. Sepulveda, S. Teixeira, U. Schuchardt, Appl. Catal., A 318 (2007) 213–217.
[11] (a) C. Orellana-Coca, U. Tornvall, D. Adlercreutz, B. Mattiasson, R. Hatti-Kaul,
Biocatal. Biotransform. 23 (2005) 431–437;
(e) J.M. Fraile, J.I. Garcia, J.A. Mayoral, E. Vispe, Appl. Catal., A 245 (2003)
363–376;
(f) J.A. Fraile, J.I. Garcia, J.A. Mayoral, E. Vispe, Appl. Catal., A 276 (2004) 113–122;
(g) M. Guidotti, C. Pirovano, N. Ravasio, B. Lazaro, J.M. Fraile, J.A. Mayoral, B. Coq,
A. Galarneau, Green Chem. 11 (2009) 1421–1427.
(b) C. Orellana-Coca, S. Camocho, D. Adlercreutz, B. Mattiasson, R. Hatti-Kaul,
Eur. J. Lipid Sci. Technol. 107 (2005) 864–870;
(c) C. Orellana-Coca, J.M. Billakanti, B. Mattiasson, R. Hatti-Kaul, J. Mol. Catal.
B: Enzym. 44 (2007) 133–137;
(d) U. Tornvall, C. Orellana-Coca, R. Hatti-Kaul, D. Adlercreutz, Enzyme Microb.
Technol. 40 (2007) 447–451;
[21] J. Dupont, R.F. de Souza, P.A.Z. Suarez, Chem. Rev. 102 (2002) 3667–3691, and
references cited therein.
[22] H. Garcia-Marin, J.C. van der Toorn, J.A. Mayoral, J.I. Garcia, I. Arends, Green
Chem. 11 (2009) 1605–1609.
[23] J.I. Garcia, H. Garcia-Marin, J.A. Mayoral, P. Perez, Green Chem. 12 (2010)
426–434.
[24] H. Garcia-Marin, J.C. van der Toorn, J.A. Mayoral, J.I. Garcia, I. Arends, J. Mol.
Catal. A 334 (2011) 83–88.
[25] M. De Torres, G. Jimenez-Osés, J.A. Mayoral, E. Pires, Bioresour. Technol. 102
(2011) 2590–2594.
[26] J.C. van der Toorn, G. Kemperman, R.A. Sheldon, I. Arends, J. Org. Chem. 74
(2009) 3085–3089.
(e) M.R.G. Klaas, S. Warwel, J. Am. Oil Chem. Soc. 73 (1996) 1453–1457.
[12] (a) J. Mlochowski, M. Brzaszcz, M. Giurg, J. Palus, H. Wojtowicz, Eur. J. Org.
Chem. (2003) 4329–4339;
(b)
G.J.
ten
Brink,
B.C.M.
Fernes,
M.C.A.
van
Vliet,
I.
Arends, R.A. Sheldon, J. Chem. Soc., Perkin Trans.
224–228.
1 (2001)
[13] (a) S.B. Said, J. Skarzewski, J. Mlochowski, Synth. Commun. 22 (1992)
1851–1862;
[27] L.S. Silbert, D. Swern, E. Siegel, J. Org. Chem. 27 (1962) 1336–1342.