Fatty ketones from the rearrangement of epoxidized vegetable oils

Article abstract of DOI:10.1016/j.apcata.2012.09.007

The rearrangement of epoxidized vegetable oils to produce fatty ketones, catalyzed by acidic resins, is disclosed. The non-microbial production of fatty ketones from epoxidized vegetable oils has not been previously reported. Some of the variables affecti

Full text of DOI:10.1016/j.apcata.2012.09.007

Applied Catalysis A: General 445–446 (2012) 346–350
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Fatty ketones from the rearrangement of epoxidized vegetable oils
Luis A. Riosa,∗, Biviana A. Llanoa, Wolfgang F. Hoelderichb
a Grupo Procesos Fisicoquímicos Aplicados, Universidad de Antioquia, Cra. 53 # 61-30 Medellín, Colombia
b Department of Chemical Technology and Heterogeneous Catalysis, RWTH Aachen University, Worringerweg 1, 52074 Aachen, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 9 May 2012
Received in revised form 31 August 2012
Accepted 3 September 2012
Available online 12 September 2012
The rearrangement of epoxidized vegetable oils to produce fatty ketones, catalyzed by acidic resins, is
disclosed. The non-microbial production of fatty ketones from epoxidized vegetable oils has not been
previously reported. Some of the variables affecting the ketone formation such as acidic strength of
the catalyst, temperature and solvent are studied in order to determine the conditions that favor the
rearrangement. Epoxidized vegetable oils can be easily transformed to the respective ketones via a
rearrangement reaction catalyzed by acidic resins. Other kind of acidic catalysts are active for this reac-
tion if their Brönsted acid sites are accessible to protonate the epoxide. Formation of ketones, from
the rearrangement of epoxidized methyl oleate, is favored in the presence of strongly acidic catalysts
and enhanced by increasing temperature. Polar-protic solvents increase the ketone yield but decrease
the ketone selectivity because they are added to the epoxide ring. The mechanism for the epoxide
rearrangement is very likely to take place through a hydride migration to the carbocation generated
in the acid-catalyzed epoxide ring opening.
Keywords:
Ketone
Epoxide
Oil
Vegetable
Acid
Resin
Rearrangement
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
phenylacetaldehydes, methyl isopropyl ketone and campholenic
aldehyde from the respective epoxides and using heteroge-
Epoxidation of vegetable oils is a commercially important reac-
tion because the epoxides obtained from these renewable raw
materials and from their alkyl esters, their transesterification prod-
ucts, have applications in such materials as plasticizers and polymer
stabilizers [1,2]. Furthermore, these epoxides can be used as inter-
mediates in the production of a variety of derivatives, because of the
high reactivity of the strained epoxide ring. For example, monoalco-
hols, diols, alkoxyalcohols, hydroxyesters, N-hydroxyalkylamides,
mercaptoalcohols, aminoalcohols, and hydroxynitriles could be
produced via epoxide ring-opening reactions with suitable reac-
tants [3]. Among these products, it was recently reported the
addition of alcohols and organic acids to epoxidized vegetable oils
to obtain derivatives that have shown good properties as biocom-
patible lubricants [4–7].
During our investigations dealing with the production of fatty
epoxides and their transformation into other valuable products
[4–9] we realized that ketones were produced as by-products,
through an epoxide rearrangement reaction. Rearrangement of
epoxides to ketones has been reported but for short-chain epoxides
such as ethylene oxide and propylene oxide [10–13]. This reaction
has been used to obtain important commercial products like
neous catalysts such as acidic zeolites [14]. Yates [15] reported a
method to oxidize olefins to the less substituted carbonyl com-
pounds, involving two steps: hydroboration with borane dimethyl
sulfide, followed by oxidation of the resulting alkylboranes with
tetrapropylammonium perruthenate and N-methylmorpholine-N-
oxide. Only short chain olefins were studied, with the unsaturation
in terminal or cyclic position. The reaction was accomplished in
ca. 4 h at room temperature. On the other hand, Deshmukh et al.
[16] reported a method to convert olefins into the corresponding
␣-bromo ketones in the presence of 1.1 equivalents each of o-
iodoxybenzoic acid and tetraethylammonium bromide. Similarly,
Maksimchuk et al. [17] established that the mesoporous metal-
organic framework MIL-101 (chromium terephthalate-based
mesoscopic metal-organic framework) is an efficient heteroge-
neous catalyst for the selective allylic oxidation of alkenes with
tert-butyl hydroperoxide. They found that the selectivity toward
␣,␤-unsaturated ketones reached 86–93% and the temperature of
the catalyst activation strongly affected the ketone yield. Wang
and Jiang [18] developed a method for palladium-catalyzed dihy-
droxylation and oxidative cleavage of olefins (monosubstituted
aromatic and aliphatic terminal alkenes, 1,2-disubstituted, and 1,1-
disubstituted olefins) with oxygen to produce 1,2-diols, aldehydes,
and ketones. Microbial production of ketones from fatty epoxides
using Rhodococcus rhodochrousand and Nocardia cholesterolicum
[19,20] has been reported. To the best of our knowledge, the
∗
Corresponding author. Tel.: +57 4 2196589; fax: +57 4 2196565.
E-mail addresses: larios@udea.edu.co, lariospfa@gmail.com (L.A. Rios).
0926-860X/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2012.09.007

L.A. Rios et al. / Applied Catalysis A: General 445–446 (2012) 346–350
347
100
90
80
70
60
50
40
30
20
10
0
non-microbial production of fatty ketones from epoxidized veg-
etable oils has not been previously disclosed. The disclosure of the
rearrangement of epoxidized vegetable oils is a very important
issue that will help to better understand previous and future
experimental results on reactions dealing with these epoxides.
For instance, there are many reports using epoxidized oils as
reactants where the reactions were monitored by measuring the
decrease in the epoxide content, via titration (oxirane value) or
infrared spectroscopy. Most of these reports have assumed that the
measured decrease in the epoxide content is due to the formation
of the desired product and do not take into account the possible
ketone formation.
This paper is devoted to this reaction, i.e., the rearrangement
of epoxidized vegetable oils to produce fatty ketones. Some of the
variables affecting the ketone formation, such as acidic strength
of the catalyst, temperature and solvent are studied in order to
determine the conditions that favor the rearrangement.
SAC13
Amberlyst15
Selectivity to ketone > 99%
0
40
80
120
160
Time (min)
200
240 280
320
360
Fig. 1. Effect of the catalyst acid strength on the ketone formation rate from epoxi-
dized methyl oleate. Epoxide/catalyst = 10 g/g, temperature = 70 ◦C.
2. Materials and methods
to the epoxide) was added. To investigate the effect of stirring speed
on the three-phase reaction system, reactions were carried out over
a range of stirring speeds (100–1000 rpm). It was observed that the
epoxides conversion increased with the stirring speed, but beyond
500 rpm there was no appreciable change in the rate of reaction.
Therefore, all the subsequent experiments were carried out under a
stirring speed of 500 rpm to ensure that there was no mass-transfer
limitation in the liquid phases. Data plotted in the accompanying
figures correspond to the means of triplicate experiments with a
relative standard deviation <5% in all cases.
2.1. Materials
Oleic acid methyl ester was used as a substrate (Fuchs Petrolub
AG, 97 wt% cis-9-octadecenoic acid methyl ester). Amberlyst 15
was purchased from Aldrich, SAC13 was kindly provided by DuPont.
Amberlyst 15 is a copolymer of styrene and divinylbenzene (DVB).
SAC13 is a composite material made of Nafion nanoparticles
entrapped in a silica matrix. Nafion itself is a copolymer of tetrafluo-
roethene and perfluor-2-(fluorosulfonylethoxy)-propylvinyl ether.
All the other reactants were from analytical grade and purchased
from Aldrich.
3. Results and discussion
The total amount of Brönsted acid sites in Amberlyst 15 and
SAC13 was 4.72 and 0.22 meq H+/g respectively, determined by ion
exchange with NaCl solution 1.0 M (10 mL per gram of catalyst) and
titrating the residual solution with NaOH 0.01 M. Surface area and
porosity were measured with a Micromeritics Gemini V employing
the BET method, prior to this measurement samples were degassed
for 4 h at 100 ◦C and 10−2 Pa.
Experiments on the rearrangement of epoxidized methyl oleate
were carried out to analyze the effect of the catalyst’s acid strength,
reaction temperature and solvent on the epoxide conversion and
ketone yield. Fig. 1 shows the results with the first variable, i.e.
acidic strength of the catalyst.
The catalyst SAC13 (13% Nafion® entrapped in silica) is quite
better than the catalyst Amberlyst15 for the ketone production.
At 70 ◦C, ca. 3.5 h are required for complete epoxide conversion
on SAC13 while Amberlyst15 needs ca. 30 h to completely convert
the epoxide at 100 ◦C. SAC13 is a very strong acid, which explains
its superior activity; Hammet acidity and NMR of mesityl oxide
(Table 1) indicate that the acidity of this catalyst is equivalent to 85%
sulfuric acid solution, while the acidity of Amberlyst 15 is equiva-
lent to 45% sulfuric acid solution. This fact is a consequence of the
fluor atom in the Nafion resin (component of SAC13), which owing
its high electronegativity increases the dissociation of the acidic
proton.
When different temperatures were studied, the results shown
in Figs. 2 and 3 were obtained for the epoxide rearrangement on
Amberlyst15 and SAC13, respectively. It is clear that the temper-
ature plays a crucial role on the ketone formation rate, which is
favored by increasing the temperature. If the ketone formation is
to be minimized then mildly acidic catalysts, like Amberlyst15, and
temperatures lower than 70 ◦C have to be used. When the cata-
lyst is strongly acidic (SAC13) the ketone formation is appreciable
even at temperatures as low as 50 ◦C. The reaction order was esti-
mated from data of Figs. 2 and 3. By plotting – ln(1 − ꢀ) vs. t, where
ꢀ is the conversion and t the reaction time, it was found that the
reaction exhibits the first order kinetics typical for intramolecular
(rearrangement) reactions. This result is supported by the high val-
ues obtained for the correlation coefficients (R2), which were higher
than 0.96 in all cases (for both catalysts at 50, 70 and 100 ◦C). Rate
constants (k) and correlation coefficients (R2) are shown in Table 2.
2.2. Characterization of reaction mixtures
Reaction mixtures were analyzed by gas chromatography on a
Hewlet Packard HP 6890 using 60 m of the slightly polar column FS-
SE54. Products were characterized by GC-Mass Spectrometry (GC
Varian 3400 CX, MS Varian Saturn 3 at 70 eV and electron ioniza-
tion), infrared spectroscopy (Nicolet Protégé 460, NaCl windows)
and 1H and 13C NMR. 1H (300 MHz) and 13C NMR (75 MHz) were
recorded in CDCl3 as the solvent using a Mercury-300 BB spectrom-
eter. For 1H NMR spectra, a total of 16 scans were performed with a
relaxation delay of 0.100 s and width of 8103.7 Hz. 320 scans were
recorded for 13C NMR spectra with a relaxation delay of 0.100 s and
width of 19607.8 Hz.
2.3. Catalytic experiments
Oleic acid methyl ester was used as substrate. Epoxidized oleic
acid methyl ester was produced following a known procedure [1].
Rearrangement reactions were carried out in round-bottom glass
reactors immersed in a heated oil bath at 57 ◦C. The agitation was
performed using a turbine impeller (2 cm radius, 500 rpm) to pre-
vent attrition of the catalyst. Solvents and substrate were kept over
a molecular sieve (UOP type 3A) to prevent water contamination
that could lead to the hydrolysis of the epoxide. Catalysts were
evacuated overnight at 120 ◦C under high vacuum and kept under
argon atmosphere. Epoxide and solvent were mixed for 15 min at
the desired temperature, then the catalyst (10% weight with respect

348
L.A. Rios et al. / Applied Catalysis A: General 445–446 (2012) 346–350
Table 1
Physicochemical properties of the acidic resins.
Amberlyst 15
SAC13
H
H
C
C
C
C
x
H2
H2
y
CF2
CF2 CF2
CF
x
n
SO3H
CF2
OCF2
OCF2CF
CF3
HSO3
C
C
H
m
H2
Structure
Composition
Styrene + 20% divinylbenzene
Tetrafluoroethene + perfluoro-2-(fluorosulfonylethoxy) propylvinyl ether
entrapped on silica
Acid strength [21]
Ho = −2.2 (Hammet acidity) NMR of mesityl oxide:
Ho ≥ −12 (Hammet acidity) NMR of mesityl oxide: ꢁı = 50–51 ppm,
ꢁı = 32.4 ppm, equivalent to 45% H2SO4
equivalent to 85% H2SO4
Acidity (meq H+/g)
Surface area (m2/g)
Pore size (nm)
4.72
51
40–80
0.22
220
10–25
Table 2
Kinetic parameters for ketone formation from epoxidized methyl oleate.
100
90
Temp. (◦C)
k (h−1
)
R2
50°C
Catalyst (%)
80
70°C
Amb. 15
Amb. 15
Amb. 15
SAC13
SAC13
SAC13
50
70
100
50
70
100
0.0012
0.0035
0.0937
0.0034
0.0229
0.1586
0.961
0.983
0.967
0.960
0.978
0.996
70
60
50
40
30
20
10
0
100°C
Selectivity to ketone > 99%
Epoxide/catalyst = 10 g/g, solvent/epoxide = 2 g/g.
chosen because they proved to be the less reactive alcohols in the
epoxide ring-opening reaction. Results are shown in Figs. 4 and 5.
In the presence of the apolar and slightly polar solvents cyclo-
hexane and toluene the formation rate of the ketone is decreased,
probably as a consequence of the dilution. The polar-aprotic sol-
vents sulfolane and nitrobenzene completely inhibit the ketone
formation. The reaction rate showed to be enhanced by the pres-
ence of the polar-protic solvents tert-butanol and neopentanol,
but in this case the main product obtained is the hydroxy-ether.
After ca. 22 h of reaction, in presence of tert-butanol, the ketone
yield began to decrease due to the formation of the trans-esterified
product in a consecutive reaction.
0
5
10
15
20
25
30
Time (h)
Fig. 2. Effect of temperature on the ketone formation rate from epoxidized methyl
oleate. Epoxide/catalyst = 10 g/g, catalyst Amberlyst15.
The influence of different solvents on the epoxide
rearrangement reaction was determined. For these studies the
solvents used were tert-butanol and neopentanol (polar-protic),
nitrobenzene and sulfolane (polar-aprotic), toluene (slightly polar)
and cyclo-hexane (apolar). Tert-butanol and neopentanol were
Regarding the mechanism for the epoxide rearrangement,
experiments showed that it did not take place via a pinacol-like
100
90
80
70
60
100
Tert-butanol
Neopentanol
80
60
40
20
0
without solvent
Toluene
cyclohexane
50
40
30
20
10
0
50°C
70°C
100°C
Selectivity to ketone > 99%
0
5
10
15
20
25
30
0
40
80
120
160
200
240 280
320
360
Time (h)
Time (min)
Fig. 4. Effect of solvents on the conversion of epoxidized methyl oleate.
Epoxide/catalyst = 10 g/g, catalyst Amberlyst 15, solvent/epoxide = 2 g/g and tem-
perature = 70 ◦C.
Fig. 3. Effect of temperature on the ketone formation rate from epoxidized methyl
oleate. Epoxide/catalyst = 10 g/g, catalyst SAC13.

L.A. Rios et al. / Applied Catalysis A: General 445–446 (2012) 346–350
349
1H NMR (300 MHz, CDCl3): ı (ppm) = 3.59 (s, 1), 2.33–2.8 (t, 6.8),
2.25–2.2 (t, 3), 1.54–1.48 (m, 4), 1.19 (m, 5, 9.10), 0.83–0.78 (t, 11).
13C NMR (75 MHz, CDCl3): ı (ppm) = 214 (7), 174.6 (2), 51.8 (1), 43.2
(6.8), 34.44 (3), 32.25–32.2 (9), 30.07–29.32 (5.10), 25.28 (4), 14.48
(11).
The IR spectrum (not shown) shows the usual vegetable-
oil transmittance bands (oleic acid methyl ester), but the band
at 3040–3010 cm−1, characteristic of ethylenic groups, is absent
which indicates the complete conversion of the unsaturations.
The IR spectrum also shows a band at 1710 cm−1 characteris-
tic of carbonyl groups. Bands at 3600–3200 cm−1, characteristic
of OH groups in hydrogen bonding are not present indicating
that the epoxide was not opened by proton-donor molecules (like
water).
60
50
40
30
20
10
0
Tert-butanol
Neopentanol
without solvent
Toluene
cyclohexane
0
5
10
15
Time (h)
20
25
30
Fig. 5. Effect of solvents on the ketone formation from epoxidized methyl oleate.
Epoxide/catalyst = 10 g/g, catalyst Amberlyst 15, solvent/epoxide = 2 g/g and tem-
perature = 70 ◦C.
4. Conclusions
Epoxidized vegetable oils can be easily transformed to the
respective ketones via a rearrangement reaction catalyzed by
acidic resins. Other kind of acidic catalysts are active for this
reaction if their Brönsted acid sites are accessible to protonate
the epoxide. Formation of ketone, from the rearrangement of
epoxidized methyl oleate, is favored by the presence of strongly
acidic catalysts like SAC13 and enhanced by increasing tem-
perature. Polar-protic solvents increase the ketone yield but
decrease the ketone selectivity because they are added to the
epoxide ring. The mechanism for the epoxide rearrangement is
very likely to take place through a hydride migration to the
carbocation generated in the acid-catalyzed epoxide ring open-
ing.
H
OH
O
O
+H
R1
C
H
C
R2
R1
C
C
R2
R1
C
H
C
H
R2
H
H
H
H
O
OH
-H
R1
C
H
C
H
R2
R1
C
H
C
R2
Fig. 6. Scheme for the epoxide rearrangement mechanism.
Acknowledgments
rearrangement because the reactions starting with the glycol
(produced after hydrolysis of the epoxidized methyl oleate) did
not show formation of ketone at all. Additionally, no glycol was
detected during the experiments of epoxide rearrangement.
Instead, a pathway involving two distinct steps is proposed. The
first step is the acid catalyzed rupture of the oxirane ring to
produce a carbocation and the second step is a hydride migration
as shown in Fig. 6. Such as mechanism has been already proposed
for the rearrangement of propene oxide and supported by high
level ab initio calculations [12,13].
Ultimately, other acidic catalysts like clays (montmorillonite
K10 and KSF/O) and mesoporous zeolite H–Y were also active for
this reaction (results not shown). Results indicate that the nec-
essary condition is the Brönsted acid sites to be accessible to
protonate the epoxide, which is the first step in the rearrangement.
Thanks are given to the German Science Foundation (Deutsche
Forschungsgemeinschaft) for its financial support of the research
project SFB 442, in which this work was developed. Sup-
port of the “Universidad de Antioquia” is also acknowl-
edged.
References
[1] J.G. Wallace, Encyclopedia of Chemical Technology, Wiley, New York, 1978,
263-264.
[2] U. Biermann, W. Friedt, S. Lang, W. Lühs, G. Machmüller, J.O. Metzger, M.
Rüsch gen. Klaas, H.J. Schäfer, M.P. Schneider, Angew. Chem. Int. Ed. 39 (2000)
2206–2224.
[3] H. Baumann, M. Bühler, H. Fochem, F. Hirsinger, H. Zoebelein, J. Falbe. Angew.
Chem. 100 (1988) 42–62.
[4] L. Rios, P. Weckes, H. Schuster, H. Haussman, W.F. Hoelderich, Appl. Catal., A
253 (2003) 487–497.
[5] L. Rios, W. Hoelderich, P. Weckes, H. Schuster, J. Synth. Lubr. 20 (2004)
289–301.
3.1. Product characterization
[6] L. Rios, H. Schuster, W. Hoelderich, P. Weckes, Appl. Catal., A 284 (2005)
155–161.
[7] H. Schuster, L. Rios, W.F. Hoelderich, P. Weckes, Appl. Catal., A 348 (2008)
266–270.
[8] L. Rios, P. Weckes, H. Schuster, W. Hoelderich, J. Catal. 232 (2005) 19–26.
[9] L. Rios, D. Echeverri, A. Franco, Appl. Catal., A 394 (2011) 132–137.
[10] J. Coxon, R. Maclagan, A. Rauk, A. Thorpe, D. Whalen, J. Am. Chem. Soc. 119
(1997) 4712–4718.
[11] J. Coxon, R. Maclagan, A. Rau, A. Thorpe, D. Whalen, Molecular gymnastics of
oxiranes in acid, in: H.S. Rzepa, J.P. Snyder, C. Leach (Eds.), Electronic Conference
on Heterocyclic Chemistry, The Royal Society of Chemistry, London, 1996 (ISBN
0 85404 894 4, on-line version).
1H NMR and 13C NMR spectra, presented below, show the for-
mation of the ketone moiety and the disappearance of the ethylenic
and epoxide moieties. Besides, the formation of the ketones was
observed in the GC–MS analysis by comparison with the mass spec-
tra available in the data bank of the equipment.
O
H2
C
H2
C
H2
C
CH3
[12] R.O.C. Norman, J.M. Coxon, Principles of Organic Synthesis, 3rd ed., Blackie,
Chapman and Hall, London, 1993, pp. 590–591.
H3CO
C
C
C
CH2
C
4
6
[13] J. Coxon, D. McDonald, Tetrahedron Lett. 29 (1988) 2575–2576.
[14] W. Hölderich, in: H. Van Bekkum, E.M. Flanigen, J.C. Jansen (Eds.), Introduction
to Zeolite Science and Practice, Elsevier, Amsterdam, 1991, pp. 655–656.
[15] M.H. Yates, Tetrahedron Lett. 38 (1997) 2813–2816.
[16] S.S. Deshmukh, K.H. Chaudhari, K.G. Akamanchi, Synlett (2011) 81–83.
[17] N.V. Maksimchuk, K.A. Kovalenko, V.P. Fedin, O.A. Kholdeeva, Adv. Synth. Catal.
352 (2010) 2943–2948.
H2
H2
H2
O
7
1
2
3
4
5
6
8
9
10 11
Octadecanoic acid, 9-oxo, methyl ester

350
L.A. Rios et al. / Applied Catalysis A: General 445–446 (2012) 346–350
[18] A. Wang, H. Jiang, J. Org. Chem. 75 (2010) 2321–2326.
[20] S. Koritala, L. Hosie, C. Hou, C. Hesseltine, M. Bagby, Appl. Microbiol. Biotechnol.
[19] S. Koritala, L. Hosie, M. Bagby, Oleic acid conversion to 10-hydroxystearic acid
by Nocardia species, in: T.H. Applewhite (Ed.), Proceedings of the World Con-
ference on Biotechnology for the fats and oils industry, American Oil Chemists
Society, Champaign, Ill, 1988, pp. 298–300.
32 (1989) 299–304.
[21] D. Farcasiu, A. Ghenciu, G. Marino, K.D. Rose, J. Am. Chem. Soc. 119 (1997)
11826–11831.