Analysis of fatty acid epoxidation by high performance liquid chromatography coupled with evaporative light scattering detection and mass spectrometry

Article abstract of DOI:10.1016/j.chemphyslip.2005.02.014

Conventionally, epoxidation of unsaturated fatty acids has been studied either with titrimetric methods or in a lengthy procedure involving derivatization followed by gas chromatography (GC). We have developed a more rapid and descriptive analysis procedure for the substances using high performance liquid chromatography (HPLC) with evaporative light scattering detection (ELSD). Chemo-enzymatic epoxidation of unsaturated fatty acids (oleic, linoleic and linolenic acid, respectively) has been performed using hydrogen peroxide and immobilized lipase from Candida antarctica (Novozym 435). The fatty acids and their epoxidation products were separated by HPLC on a C-18 reversed-phase column using methanol-water containing 0.05% acetic acid as mobile phase. The method facilitated the simultaneous determination of fatty acids and epoxides differing from each other in the number of epoxide rings, the degree of unsaturation and the position of the epoxide rings and double bonds. An important aspect of the method development was the use of electrospray ionization and tandem mass spectrometry to confirm the structure of the epoxide products. It is suggested that the HPLC method, providing more information about the kind and concentration of fatty acids and their epoxides, represents a powerful complement to the existing methods for monitoring epoxidation processes on fatty acids.

Full text of DOI:10.1016/j.chemphyslip.2005.02.014

Chemistry and Physics of Lipids 135 (2005) 189–199
Analysis of fatty acid epoxidation by high performance liquid
chromatography coupled with evaporative light scattering
detection and mass spectrometry
Cecilia Orellana-Coca, Dietlind Adlercreutz∗, Maria M. Andersson,
Bo Mattiasson, Rajni Hatti-Kaul
Department of Biotechnology, Center for Chemistry & Chemical Engineering, Lund University, Box 124, S-221 00 Lund, Sweden
Received 25 May 2004; received in revised form 17 February 2005; accepted 23 February 2005
Available online 14 April 2005
Abstract
Conventionally, epoxidation of unsaturated fatty acids has been studied either with titrimetric methods or in a lengthy procedure
involving derivatization followed by gas chromatography (GC). We have developed a more rapid and descriptive analysis
procedure for the substances using high performance liquid chromatography (HPLC) with evaporative light scattering detection
(ELSD). Chemo-enzymatic epoxidation of unsaturated fatty acids (oleic, linoleic and linolenic acid, respectively) has been
performed using hydrogen peroxide and immobilized lipase from Candida antarctica (Novozym 435). The fatty acids and their
epoxidation products were separated by HPLC on a C-18 reversed-phase column using methanol–water containing 0.05% acetic
acid as mobile phase. The method facilitated the simultaneous determination of fatty acids and epoxides differing from each other
in the number of epoxide rings, the degree of unsaturation and the position of the epoxide rings and double bonds. An important
aspect of the method development was the use of electrospray ionization and tandem mass spectrometry to confirm the structure
of the epoxide products. It is suggested that the HPLC method, providing more information about the kind and concentration of
fatty acids and their epoxides, represents a powerful complement to the existing methods for monitoring epoxidation processes
on fatty acids.
© 2005 Elsevier Ireland Ltd. All rights reserved.
Keywords: Epoxidation; Unsaturated fatty acids; HPLC analysis; Mass spectrometry; Lipase
1. Introduction
The utilization of oils and fats for the production of
chemicals and materials that can act as a replacement
for those derived from petroleum is becoming increas-
∗
Corresponding author. Tel.: +46 46 222 7363;
fax: +46 46 222 4713.
E-mail address: dietlind.adlercreutz@biotek.lu.se
ingly important (Biermann et al., 2000). Vegetable oils
(D. Adlercreutz).
0009-3084/$ – see front matter © 2005 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.chemphyslip.2005.02.014

190
C. Orellana-Coca et al. / Chemistry and Physics of Lipids 135 (2005) 189–199
and their unsaturated fatty acids have, for instance,
been converted to epoxides (Klaas and Warwel, 1999;
Ruesch gen. Klaas and Warwel, 1996) that are useful
building blocks in organic synthesis as they participate
in numerous reactions owing to their highly reactive
oxirane ring. Oleochemical internal epoxides are
industrially applied as PVC-stabilizers, plasticizers,
as reactive diluents for paints and as intermediates
for polyurethane–polyol production (Ramirez-de-
Arellano-Aburto et al., 2002; Wilkes et al., 2004).
Epoxidized vegetable oils and their derivatives are
also suitable as cross-linkers in environmentally
friendly coatings, where they serve as a substitute
for triglycidyl isocyanurate, a putative mutagenic
substance (Overeem et al., 1999).
Fatty epoxides are currently produced on an indus-
trial scale by the Prileshajev epoxidation reaction in
which a peracid is used for oxygen transfer to the dou-
ble bonds (Swern, 1947). (Warwel and Klaas, 1995)
have earlier described a selective epoxidation process,
which involves the use of a lipase to catalyze the for-
mation of peracid from H2O2 and the corresponding
fatty acid. The reaction scheme for the lipase-mediated
epoxidation of oleic acids is presented in Fig. 1.
A combination of two titrimetric methods is com-
monly employed to monitor the epoxidation of fatty
acids and oils (Klaas and Warwel, 1999). These include
the determination of iodine value, which is a measure of
the number of double bonds (Wijs, 1929), and oxirane
number, which is a measure of the amount of epoxide
oxygen (Jay, 1964). The epoxidation process should go
hand-in-hand with a decrease in the iodine value and
an increase in the oxirane number. The degree of epox-
idation can in principle be calculated from these two
values. Generally, the quality of the epoxidized fatty
acid or oil is better, the higher the oxirane number and
the lower the iodine value (Carlson and Chang, 1985;
Hang and Yang, 1999).
tain more detailed information on the composition of
epoxidized fatty acids and oils, these have been con-
verted to methyl esters and then been analyzed by GC
in combination with FID and/or mass spectrometric de-
tection (Overeem et al., 1999; Ruesch gen. Klaas and
Warwel, 1996; Warwel and Klaas, 1995). However, as
samples are exposed to high temperatures, derivatiza-
tion of the fatty epoxides cannot be excluded com-
pletely. Methyl esters of epoxidized fatty acids have
also been analyzed by HPLC with evaporative light
scattering detection (ELSD) using two symmetry C-18
reversed-phase columns connected in series and ace-
tonitrile and water as the mobile phase (Piazza et al.,
2003). HPLC-electron impact mass spectrometric (EI-
MS) detection and atmospheric pressure chemical ion-
ization mass spectrometric (APIC-MS) detection were
used for characterization of the fatty epoxides.
In this report we present a simple HPLC method
with ELSD and electrospray ionization (ESI) iden-
tification, which provides simultaneous analysis and
structural identification of both the fatty acids and the
corresponding epoxides, including the regio-isomers.
In contrast to the other methods, derivatization of the
free fatty acids prior to analysis was not required. Ex-
ternal standards were used for calibration of the ELSD,
and the synthesis and purification of these standards is
also described. Finally, the method has been applied
for monitoring the chemo-enzymatic process of epox-
idation of fatty acids with different degrees of unsatu-
ration.
2. Materials and methods
2.1. Materials
Candida antarctica lipase (Novozym®435) was a
gift from Novozymes (Bagsvaerd, Denmark). Oleic
acid, linoleic acid, linolenic acid, stearic acid, palmitic
acid, heptadecanoic acid and epoxy stearic acid were
purchased from Sigma–Aldrich (St. Louis, USA).
Hydrogen peroxide (30%, w/w) in water, glacial
acetic acid, toluene, sulphuric acid, ethanol, para-
anisaldehyde, silica gel 60 (0.04–0.063 ␮m), TLC alu-
minium sheets, silica gel 60 F254, heptane, methanol of
HPLC grade and ethyl ether (SeccoSolv®) were pro-
cured from Merck (Darmstadt, Germany). The water
used was purified with a Milli-Q system from Milli-
This analysis, however, is afflicted with an error,
arising from the fact that the iodine value is typically
strongly dependent on the method by which it is deter-
mined (Kyriakidis and Katsiloulis, 2000). Moreover,
these titration methods are inadequate for providing
information on the fatty acid composition and the de-
gree of epoxidation of individual fatty acids, charac-
teristics that can be expected to have a large impact
on the applicability of the epoxide product (Klaas and
Warwel, 1999; Overeem et al., 1999). In order to ob-

C. Orellana-Coca et al. / Chemistry and Physics of Lipids 135 (2005) 189–199
191
Fig. 1. Reaction scheme for the chemo-enzymatic epoxidation of oleic acid. Oleic acid is first converted by the lipase to yield the corresponding
peracid, which then epoxidizes the double bond of oleic acid in a predominantly intermolecular reaction (Warwel and Klaas, 1995).
pore (Billerica, USA). All the reagents were of analyt-
ical grade.
at room temperature. The commercially available pure
fatty acids were used as standards; the epoxide stan-
dards, except for epoxy stearic acid, had to be prepared
since they are not commercially available in pure form.
2.2. HPLC analysis of fatty acids and epoxides
HPLC analyses were carried out on a LaChrom
HPLC system, equipped with an L-2500 pump used
in low gradient mode, an L-2400 automatic sample
injector, a D-7000 Interface (all from Merck Hitachi)
and an evaporative light scattering detector , model 500
with a laser diode light source from Alltech (Deerfield,
IL). The ELSD was operated at 87 ◦C with a gas flow of
2.37 standard litres/min. Air was used as nebulizer gas.
The fatty acids and their epoxidation products were
separated on a Lichrospher RP-C18 column (5 ␮m,
250 mm × 4 mm) with an RP-C18 guard cartridge
(5 ␮m, 4 mm × 4 mm) from Merck (Darmstadt, Ger-
many). The injection volume was 20 ␮l. The mobile
phase solvents were: solvent A, methanol–acetic acid
(99.95:0.05, v/v) and solvent B, water–acetic acid
(99.95:0.05, v/v). The gradient program employed was
as follows: segment 1—isocratic 75% A, 25% B for
12 min; segment 2—linear gradient from initial condi-
tions to 90% A; 10% B for 8 min; segment 3—linear
gradient to 98% A; 2% B for 15 min; segment 4—linear
gradient from 98% A; 2% B to initial condition 75%
A; 25% B for 2 min; and segment 5—isocratic 75%
A, 25% B for 3 min. The flow rate used was 1 ml/min
2.3. Lipase-mediated epoxidation of fatty acids:
monitoring by HPLC
The reaction medium consisted of a 0.5 M solution
of fatty acid (oleic acid, linoleic acid and linolenic acid,
respectively) in 5 ml of toluene (10 ml for linoleic acid)
and 100 mg immobilized C. antarctica lipase. Hydro-
gen peroxide (30%, w/w) was added stepwise in order
to prevent the enzyme from being inactivated. The rate
of addition was adjusted for every fatty acid substrate
(oleic acid: 30 ␮l every 15 min during 2.5 h; linoleic
acid: 50 ␮l every 15 min during 6 h; and linolenic acid:
40 ␮l every 15 min during 5 h). The reaction mixtures
were stirred with a magnetic stirrer bar. At appropri-
ate time intervals (about every hour), the stirring was
stopped and the two phases were allowed to separate
for about 1 min. A 50 ␮l (or 100 ␮l) sample was then
withdrawn from the organic phase with a syringe and
diluted 1:100 (or 1:200) with methanol before being
analyzed by HPLC. Once the sample was withdrawn,
no further conversion occurred, since the enzyme, be-
ing immobilized on a solid support, remained in the
reaction vessel.

192
C. Orellana-Coca et al. / Chemistry and Physics of Lipids 135 (2005) 189–199
2.4. Preparation of epoxide standards
dried by evaporation and analyzed by MS. Di-epoxy
octadecanoic acid was further purified by crystalliza-
tion to yield two diastereomers.
The fatty acid epoxides, except epoxystearic acid,
were prepared by chemo-enzymatic epoxidation of
the fatty acids using the reaction conditions described
above. The reaction was terminated after 21.8 and
10.2 h for linoleic and linolenic acid, respectively. Di-
ethyl ether (30 ml) was added to the reaction mixture,
and the enzyme preparation was filtered off to termi-
nate the reaction. If necessary, the reaction mixture was
stored at −20 ◦C before further processing the next
working day. The organic layer, containing the epox-
idized fatty acids, was washed with water to remove
traces of H2O2, and concentrated in a rotary evapora-
tor. The oily residue obtained after evaporation, con-
sisting of the epoxides and possibly traces of solvent
and water, was then purified by flash chromatography.
The purity and identity of the epoxides were confirmed
by thin layer chromatography (TLC) and electrospray
MS.
2.6. Thin layer chromatography analysis
TLC of the fractions obtained from flash chro-
matography was done on aluminium sheets coated
with silica gel 60 F254. The samples were spot-
ted on the sheets using a 5 ␮l micro glass pipette
and eluted with a mobile phase system compris-
ing heptane, ethyl ether and acetic acid (50:50:1).
After elution, the sheets were sprayed with para-
anisaldehyde reagent (ethanol/sulphuric acid/para-
anisaldehyde/acetic acid (338:12.5:9.2:3.75, v/v/v/v),
dried and heated at 105 ◦C. TLC provided separation
of fatty acids and their mono-, di- and tri-epoxides, with
the Rf values being 0.5, 0.4, 0.2 and 0.1, respectively.
Saturated and unsaturated fatty acids containing the
same number of epoxy groups were not distinguished.
2.5. Flash chromatography
2.7. Acquisition of MS and MS/MS data
The flash chromatography equipment, fabricated
in-house at the glass workshop of the Engineer-
ing Faculty of Lund, consisted of a glass column
(290 mm × 20 mm), a 500 ml solvent reservoir and a
flow controller valve. Air, supplied at a pressure of
0.2 bar, was used to force the eluent through the col-
umn. Thirty grams of silica gel 60 (0.04–0.063 ␮m)
was suspended in the eluent and poured into the col-
umn to a height of 200 mm. The concentrated epox-
ide sample (300 mg) obtained above was applied on
top of the silica column, and the sample components
were then eluted from the sorbent with the mobile
phase at a flow rate of approximately 17 ml/min. The
eluent consisted of heptane–diethyl ether–acetic acid
75:25:1 (v/v/v) to elute the mono-epoxidation products
from linoleic and linolenic acid, respectively. The di-
epoxidation products (di-epoxy octadecanoic acid and
di-epoxy octadecenoic acid, respectively) were eluted
with heptane–diethyl ether–acetic acid 50:50:1 (v/v/v),
and tri-epoxy octadecanoic acid was eluted with a
mixture of heptane–diethyl ether–acetic acid 40:60:1
(v/v/v). The effluent from the column was collected in
10 ml fractions. From these fractions, samples of about
5 ␮l were withdrawn foranalysisby TLC. The fractions
showing a single band on the TLC plate were pooled,
All the experiments were conducted on a hybrid QS-
TAR Pulsar quadrupole TOF mass spectrometer (Ap-
plied Biosystems Sciex Instruments, Toronto, Canada)
equipped with an electrospray ionization source. Initial
separation was achieved on a Perkin Elmer Series 200
capillary LC system. Samples were prepared by dis-
solving the products from each reaction in methanol
and filtering through 0.2 ␮m GHP membrane filters
from Gelman Laboratory, supplied by Merck (Darm-
stadt, Germany). Samples of 10 ␮l were injected and
separated on a Lichrospher RP-C18 column at a flow
rate of 1 ml/min, using the same separation parame-
ters as described in Section 2.2 except that formic acid
(0.05%) was used in the mobile phase instead of acetic
acid. Prior to entering the mass spectrometer, the mo-
bile phase flow was split (1:10) with a Valco valve
placed on the instrument.
MS and MS/MS spectra were recorded in the mass
range of m/z 50–400 amu with the electroionspray op-
erating at 350 ◦C and using negative mode (despite
the acidic pH of the eluent phase) with an electro-
spray voltage of −4400 V and the Q1 resolution set
to LOW. Nitrogen was used as curtain gas at 20.0 psi
and as nebulizing and drying gas at 20 and 35 psi, re-
spectively. Optimized compound parameters proved to

C. Orellana-Coca et al. / Chemistry and Physics of Lipids 135 (2005) 189–199
193
be: declustering potential, −65 V; focusing potential,
−245 V and collision energy, −40 eV. Argon with a
purity of 99.996% was used as the collision gas. Pre-
cursor ions were chosen with m/z > 250 and exceed-
ing 10 counts. Data were acquired and processed with
AnalystTM QS software.
can be distinguished by MS and MS/MS. Table 1 gives
the retention times for the epoxide products along with
the MS and tandem MS data used for the identification
of the epoxides.
As expected, the polarity of the compounds had a
major influence on their retention time. The unsatu-
rated fatty acids eluted before saturated ones with the
same number of epoxy groups, while highly epoxidized
fatty acids eluted before those containing fewer num-
bers of epoxy groups. The retention time was also in-
fluenced by the position of the double bond and epoxy
group, though to a minor extent; regio-isomers thus
eluted close to each other.
2.8. Melting point determination
The melting points of the two diastereomers 9-10,
12-13 diepoxy octadecanoic acid were measured in a
melting point device from KIFA (Reichert, Austria).
Peracids, the reactive intermediates in the lipase-
mediated epoxidation of fatty acids, comprising <2%
of the total fatty acids (Warwel and Klaas, 1995), were
not detected in the concentration range investigated.
3. Results and discussion
3.1. HPLC analysis
Vegetable oils contain a number of saturated and un-
saturated fatty acids, the concentrations of which vary
depending on the plant source. It is only the unsaturated
fatty acids that undergo epoxidation. The epoxide prod-
ucts formed may differ in the number of epoxy groups,
the degree of unsaturation and in the position of epoxy
groups and double bonds. Hence, a method that can be
employed for simultaneous analysis for the variety of
components present in a reaction mixture would pro-
vide valuable information on the epoxidation process.
A prerequisite for evaluating the scope and sensi-
tivity of the analytical method was the preparation of
standard compounds due to their lack of availability
from commercial sources. The epoxide standards were
produced by chemo-enzymatic epoxidation, since this
method is known to give fewer unwanted by-products
than any other method (Klaas and Warwel, 1999).
All the fatty acids and their various epoxidation
products, including regio-isomers, were separated us-
ing HPLC on a reverse phase column and detected by
an ELSD (Fig. 2). The saturated fatty acids, palmitic
(C16:0) and stearic (C18:0) acid were eluted after
30.7 and 34.2 min, respectively. Heptadecanoic acid
(C17:0), eluting between palmitic and stearic acid, is
not found in nature and might thus serve as an inter-
nal standard. The unsaturated fatty acids, oleic (C18:1),
linoleic (C18:2) and linolenic (C18:3) acid eluted af-
ter 30.7, 28.1 and 26.0 min, respectively. Palmitic and
oleic acid thus co-eluted and their separation based on
retention time was not possible. But the compounds
3.2. Identification of the products
Productsformed uponpartial epoxidationof linoleic
and linolenic acid (e.g., mono-epoxy octadecanoic
acid, mono-epoxy octadidecenoic acid and di-epoxy
octadecenoic acid) exhibit different regio-isomeric
forms, hard to distinguish by ELSD alone. The identi-
ties of these products were further confirmed by mass
spectrometric and tandem mass spectrometric detec-
tion. The ion source was operated in the negative mode
to produce [M − H]− precursor ions, which are prone to
charge-remotefragmentation(CRF)(ChengandGross,
2000). CRF of fatty acids occurs as H2 elimination
along the acyl chain to give both an alkene (invisible in
the mass spectrum) and a charged species possessing
the carboxylic group (Cheng et al., 1998). Accordingly,
saturated fatty acids fragment into an array of prod-
uct ions being separated by 14 mass units (57, 71, 85,
etc.). Interruption of the alkyl chain by double bonds
or epoxy groups causes a deviation from that peak pat-
tern. Moreover, the specific fragment obtained after
the interruption provides information on the kind of
substitution. ESI–MS/MS spectra obtained upon CRF
are usually straightforward to interpret and have been
widely used for structural determination of fatty acids
and related compounds (Hsu and Turk, 2004; Perret et
al., 2004). In contrast, positively charged ions fragment
in a more complex manner, which makes structural in-
formation more difficult to extract (Cheng and Gross,
2000).

194
C. Orellana-Coca et al. / Chemistry and Physics of Lipids 135 (2005) 189–199
Table 1
Retention time of fatty acid epoxidation products and the MS and MS/MS data used for their identification
Fatty acid substrate
Retention time (ELSD) of
epoxidation product(s)
Precursor ion
No. of epoxy
groups
Major fragments
m/z > 155
Epoxidation product
C18:1/1(9)
[M − H]−
C18:1
C18:2
23.6
297
1
171
10.4
21.2
21.5
311
295
295
2
1
1
171, 199
195
171
C18:2/2(9,12)
C18:2/1(12)
C18:2/1(9)
C18:3
4.2
6.3
8.1
325
309
309
309
293
293
293
3
2
2
2
1
1
1
171, 199
C18:3/3(9,12,15)
C18:3/2 (9,15)
C18:3/2 (9,12)
C18:3/2(12,15)
C18:3/1(9)
171, 185, 251
171, 199, 211
183, 195, 211, 223
171
8.5
18.1
18.5
19.0
235
C18:3/1(15)
C18:3/1(12)
183,195, 211, 223
Abbreviations for the epoxidation products are the same as in Fig. 2, with the position of the epoxy groups being given in parentheses.
The total ions chromatogram (TIC) from an
HPLC–ESI–MS/MS experiment involving the epoxi-
dation products of linolenic acid is shown in Fig. 3.
Peaks around 18–19 min corresponded to precursor
ions with a mass to charge ratio (m/z) of 293 amu,
which is 16 units more than the m/z of the linolenic
acid precursor ion. It can thus be concluded that
thesepeaksoriginatefrommono-epoxyoctadidecenoic
regio-isomers. The peaks around 9 min originate from
the precursor ion with an m/z of 309 amu, thus repre-
Fig. 2. HPLC chromatogram of fatty acids and fatty acids epoxides, obtained by chemo-enzymatic epoxidation of oleic, linoleic and linolenic
acid, respectively. Mobile phase: methanol/water containing 0.05% acetic acid, as described in Section 2.2, flow rate 1 ml/min, temperature
ambient, injection volume 20 ␮l. Peaks corresponding to commercial external standards: epoxy stearic acid (C18:1/1), linolenic acid (C18:3),
linoleic acid (C18:2), oleic acid (C18:1), heptadecanoic acid (C17:0) and stearic acid (C18:0). Peaks corresponding to epoxy fatty acids,
prepared by chemo-enzymatic epoxidation and identified by HPLC tandem mass spectrometry: 9-10, 12-13, 15-16 tri-epoxy octadecanoic acid
(C18:3/3); 9-10, 15-16 di-epoxy 12 octadecenoic acid (C18:3/2a); 9-10, 12-13 di-epoxy 15 octadecenoic acid (C18:3/2b); 12-13, 15-16 di-epoxy
9 octadecenoic acid (C18:3/2c); 9-10, 12-13 di-epoxy octadecanoic acid (C18:2/2); 9-10 mono-epoxy, 12, 15 octadidecenoic acid (C18:3/1a),
15-16 mono-epoxy 9, 12 octadidecenoic acid (C18:3/1b), C18:3/1 c corresponds to 12-13 mono-epoxy 9,15 octadidecenoic acid; C18:2/1a:
12-13 mono-epoxy 9 octadecenoic acid and C18:2/1b: 9-10 mono-epoxy 12 octadecenoic acid. The concentration of the analytes was between
0.25 mM (di-epoxy octadecanoic acid) and 2.4 mM (stearic acid).

C. Orellana-Coca et al. / Chemistry and Physics of Lipids 135 (2005) 189–199
195
Fig. 3. Total ion chromatogram of the epoxidation products of linolenic acid where each peak is represented with negatively charged molecular
ion [M − H]−1. The three peaks with m/z 293 amu represent the mono-epoxy octadidecenoic acid regio-isomers, the peaks 309 represent the
di-epoxy octadecenoic acid regio-isomers, and peak 325 represents 9-10, 12-13, 15-16 tri-epoxy octadecanoic acid. The experimental details
are provided in the text.
senting the di-epoxy octadecenoic acid regio-isomers
while the peak around 4 min, with an m/z of 315 amu
represents 9-10, 12-13, 15-16 tri epoxy octadecanoic
acid.
and C10 and peaks with m/z of 211 and 223 amu were
characteristic for fragments containing oxygen and are
shown in the MS/MS pattern of 12-13 mono-epoxy
9,15 octadidecenoic acid (C18:3/1 c; Fig. 4c). Spectra
obtained with other epoxy fatty acids were interpreted
in a similar way. The ESI–MS and ESI–MS/MS data
used for the determination of their molecular structure
are given in Table 1.
Fig. 4a–c shows the ESI–MS/MS spectra of the
three regio-isomers of mono-epoxy octadidecenoic
acid, having retention times on HPLC of 18.3, 18.6
and 19.0 min, respectively. As expected, it was found
that fragments comprising the first nine carbon atoms
(m/z ≤ 155 amu;e.g., 71, 97and127 amu)werenotuse-
ful for structural elucidation. Larger fragments, on the
other hand, contained either oxygen or double bonds
and could thus be used to differentiate between dif-
ferent regio-isomers. The peak with m/z of 171 amu
for instance was characteristic for the epoxy group
being between C9 and C10. It was thus concluded
that ESI–MS/MS spectra recorded at 18.3 min origi-
nate from 9-10 mono-epoxy 12, 15 octadidecenoic acid
(18:3/1a; Fig. 4a). The peak with m/z of 235 amu, on the
otherhand, wascharacteristicforafragmentcontaining
two double bonds as shown in the MS/MS spectra of
15-16 mono-epoxy 9, 12 octadidecenoic acid (18:3/1 b;
Fig. 4b). Peakswithm/zof183and195 amuwerefound
in epoxidation products with a double bond between C9
3.3. Calibration of the ELSD
Standards for the epoxides of linoleic and linolenic
acid were prepared by chemo-enzymatic epoxidation
and purified by flash chromatography. Purification of
the reaction mixture allowed a good resolution of
mono-, di- and tri-epoxides, but unfortunately not good
resolution of the regio-isomers. To solve this problem,
it was assumed that regio-isomers give the same re-
sponse. The validity of this assumption is supported
by several studies, showing that the response of the
ELSD is the same for structurally similar analytes; for
a review see (Young and Dolan, 2004). The calibra-
tion curves for each fatty acid and epoxide were made
with 10 concentrations in a range of 0.5–5 mM. For

196
C. Orellana-Coca et al. / Chemistry and Physics of Lipids 135 (2005) 189–199
Fig. 4. MS/MS spectrum of epoxidized products of linolenic acid: (a) 9-10 monoepoxy 12, 15 octadidecenoic acid, (b) 15 monoepoxy 9,12
octadidecenoic acid and (c) 12-13 monoepoxy 9,15 octadidecenoic acid. The experimental details are provided in the text.

C. Orellana-Coca et al. / Chemistry and Physics of Lipids 135 (2005) 189–199
197
Fig. 4. (Continued).
regio-isomers, the total area covered by the peaks of
3.4. Monitoring of epoxidation reactions by HPLC
the isomer products was taken into consideration. The
detector response could be described by a quadratic
equation. The non-linear response is quite typical for
the ELSD (Heron et al., 2004). The equation for each
of the substances and the corresponding correlation co-
efficients are shown in Table 2. The detection limits,
defined as three times the noise levels, are also given
in Table 2.
The applicability of the HPLC method was then
evaluated for monitoring the lipase-mediated epoxi-
dation of unsaturated fatty acids, by using the ELSD.
The progress of the epoxidation of oleic acid is shown
in Fig. 5. The reaction yielded epoxy stearic acid
as the only product and was complete within 10 h.
The epoxidation of linoleic acid yielded first mono-
Table 2
Standard curve constants from the calibration curve
Compound
a0
a1
a2
R
Detection limit (␮g)
C18:3/3
C18:3/2
C18:2/2
C18:3/1
C18:2/1
C18:1/1
C18:3
C18:2
C18:1
C18:0
2.076E−1
3.743E
2.555E
5.423E
4.172E
1.914E
3.750E
3.109E
3.782E
5.054E
1.286E−6
2.318E
8.807E
9.908E
5.627E
5.044E
8.044E
6.734E
6.164E
5.304E
−6.445E−14
−1.162E−13
−3.254E−14
−2.005E−14
−1.958E−5
−2.059E−15
−2.542E−14
−1.637E−14
−1.587E−14
−1.303E−14
0.997
0.987
0.995
0.996
0.997
0.997
0.993
0.993
0.989
0.997
2.1
2.9
2.8
2.3
4.1
1.9
3.9
3.4
3.1
1.2
The data were fitted to a quadratic equation C = a0 + a1X + a2X2, where C represents the concentration measured in millimolar, a0, a1 and a2 are
constants for the equations and X is the peak area.

198
C. Orellana-Coca et al. / Chemistry and Physics of Lipids 135 (2005) 189–199
Fig. 7. Progress of the chemo-enzymatic epoxidation of linolenic
acid monitored by HPLC. The reaction mixture consisted of 5 ml of
a 0.5 M solution of linolenic acid in toluene and 102 mg of Novozym
435, to which hydrogen peroxide (30%, w/w) was added in 40 ␮l
aliquots every 15 min during 5.25 h. Samples of 100 ␮l were with-
drawn every hour and diluted 100 times in methanol. (ꢀ) Linolenic
acid, (ꢁ) monoepoxy octadidecenoic acid, (ꢂ) diepoxy octadecenoic
acid and (ꢀ) triepoxy ocatadecanoic acid. Isomers are not distin-
guished. Information concerning the isomers is provided in the text.
Fig. 5. Progress of the chemo-enzymatic epoxidation of oleic acid
monitored by HPLC. The reaction mixture consisted of 5 ml of a
0.5 M solution of oleic acid in toluene and 100 mg of Novozym 435,
to which hydrogen peroxide (30%, w/w) was added in 30 ␮l aliquots
every 15 min during 2.5 h. Samples of 50 ␮l were withdrawn and
diluted 200 times in methanol. (ꢀ) Oeic acid and (ꢁ) epoxy stearic
acid.
epoxy octadecenoic acid, which was then further con-
verted to di-epoxy octadecanoic acid (Fig. 6). Di-
epoxy octadecanoic acid was the major product (al-
most 60%) after 22 h reaction time. The HPLC chro-
matogram revealed that the mono-epoxide of linoleic
acid consisted of two regio-isomers, one with the epoxy
group positioned between C9 and C10 and another
with the epoxy group positioned between C12 and
C13. There was no preference for the formation of ei-
ther isomer. Also di-epoxy octadecanoic acid eluted
in a double peak. The two peaks were found to be
two diastereomers, one with a melting point of 78 ◦C
and another with a melting point of 36 ◦C. The iso-
mer with a melting point of 78 ◦C was preferentially
formed.
The progress of the epoxidation of linolenic acid
is shown in Fig. 7. The reaction yielded first mono-
epoxy octadidecenoic acid (three regio-isomers), then
di-epoxy octadecenoic acid (three regio-isomers) and
finally tri-epoxy octadecanoic acid. The final prod-
uct contained about 30% of tri-epoxy octadecanoic
acid, ∼50% di-epoxy octadecenoic acid and ∼20%
of mono-epoxy octadidecenoic acid (Fig. 7). The
HPLC chromatograms revealed that in linolenic acid
the positions 9-10 and 15-16 were more easily
epoxidized than position 12-13, as seen in both
the mono- and di-epoxy product intensities. The
total concentration of the di-epoxy octadecenoic
acid formed was constituted by 75% 9-10, 15-16
di-epoxy 12-octadecenoic acid, 18% of 9-10, 12-
13 di-epoxy 15-octadecenoic acid and 7% of 12-
13, 15-16 di-epoxy 9-octadecenoic acid. Interest-
ingly, a similar observation has been made in a
study dealing with peroxygenase-catalyzed epoxi-
dation of unsaturated fatty acids (Piazza et al.,
2003).
Fig. 6. Progress of the chemo-enzymatic epoxidation of linoleic acid
monitored by HPLC. The reaction mixture consisted of 10 ml of a
0.5 M solution of linoleic acid in toluene and 100 mg of Novozym
435, to which hydrogen peroxide (30%, w/w) was added in 50 ␮l
aliquots every 15 min during 6 h. (ꢀ) Linoleic acid, (ꢁ) monoepoxy
octadecenoic acid and (ꢂ) diepoxy octadecanoic acid. Isomers are
not distinguished. Information concerning the isomers is provided in
the text.

C. Orellana-Coca et al. / Chemistry and Physics of Lipids 135 (2005) 189–199
199
4. Conclusion
evaporative light scattering detector in liquid chromatography-
evaporative light-scattering detection. J. Chromatogr. A 1035,
221–225.
The study shows clearly that HPLC in combination
with ELSD and ESI–MS/MS detection provides a great
deal of information on the components formed during
the epoxidation of fatty acids found in vegetable oils.
The method is applicable for monitoring of epoxidation
processes on fatty acids in general and can also be used
for analyzing epoxidized vegetable oils of different ori-
gin. Studies on the effect of the molecular composition
of the epoxidized products on their applicability in sur-
face coatings are currently under progress.
Hsu, F.-F., Turk, J., 2004. Studies on sulfatides by quadrupole ion-
trap mass spectrometry with electrospray ionization: structural
characterization and the fragmentation processes that include an
unusual internal galactose residue loss and the classical charge-
remote fragmentation. J. Am. Soc. Mass Spectrom. 15, 536–
546.
Jay, R.R., 1964. Direct titration of epoxy compounds and aziridines.
Anal. Chem. 36, 667–668.
Klaas, M.R., Warwel, S., 1999. Complete and partial epoxidation of
plant oils by lipase-catalyzed perhydrolysis. Ind. Crop Prod. 9,
125–132.
Kyriakidis, N.B., Katsiloulis, T., 2000. Calculation of iodine value
from measurements of fatty acid methyl esters of some oils: com-
parison with the relevant american oil chemists’ society method.
J. Am. Oil Chem. Soc. 77, 1235–1238.
Acknowledgments
Overeem, A., Buisman, G.J.H., Derksen, J.T.P., Cuperus, F.P., Mol-
hoek, L., Grisnich, W., Goemans, C., 1999. Seed oils rich in
linolenic acid as renewable feedstock for environment-friendly
crosslinkers in powder coatings. Ind. Crop Prod. 10, 157–165.
Perret, D., Gentili, A., Marchese, S., Sergi, M., Caporossi, L., 2004.
Determination of free fatty acids in chocolate by liquid chro-
matography with tandem mass spectrometry. Rapid Commun.
Mass Spectrom. 18, 1989–1994.
Piazza, G.J., Nunez, A., Foglia, T.A., 2003. Epoxidation of fatty
acids, fatty methyl esters, and alkenes by immobilized oat seed
peroxygenase. J. Mol. Catal. B: Enzym. 21, 143–151.
Ramirez-de-Arellano-Aburto, N., Cohen-Barki, A., Cruz-Gomez,
M.J., 2002. Improved process for the production of oleochemical
polyols from vegetable oils for polyurethane foams preparation,
US 2001-989464, (Mex.), 5 pp.
The authors are grateful to Dr. Curt T. Reimann, at
the Department of Analytical Chemistry (Lund Uni-
versity) for his valuable comments on the manuscript.
The Swedish International Development Agency/The
Swedish Agency for Research Collaboration with De-
veloping countries, and The Foundation for Strategic
Environmental Research (Mistra) are acknowledged
for financial support.
References
Biermann, U., Friedt, W., Lang, S., Luhs, W., Machmuller, G., Met-
zger, J.O., Klaas, M.R., Schafer, H.J., Schneider, M.P., 2000. New
syntheses with oils and fats as renewable raw materials for the
chemical industry. Angew. Chem., Int. Ed. 39, 2206–2224.
Carlson, K.D., Chang, S.P., 1985. Chemical epoxidation of a natural
unsaturated epoxy seed oil from vernonia galamensis and a look
at epoxy oil markets. J. Am. Oil Chem. Soc. 62, 934–939.
Cheng, C., Gross, M.L., 2000. Applications and mechanisms of
charge-remote fragmentation. Mass Spectrom. Rev. 19, 398–420.
Cheng, C., Pittenauer, E., Gross, M., 1998. Charge-remote fragmen-
tations are energy-dependent processes. J. Am. Soc. Mass Spec-
trom. 9, 840–844.
Ruesch gen. Klaas, M., Warwel, S., 1996. Chemoenzymic epoxida-
tion of unsaturated fatty acid esters and plant oils. J. Am. Oil
Chem. Soc. 73, 1453–1457.
Swern, D., 1947. Electronic interpretation of the reaction of olefins
with organic per acids. J. Am. Chem. Soc. 69, 1692–1698.
Warwel, S., Klaas, M.R., 1995. Chemoenzymic epoxidation of un-
saturated carboxylic acids. J. Mol. Catal. B: Enzym. 1, 29–35.
Wijs, J.J.A., 1929. The Wijs method as the standard for iodine ab-
sorption. Analyst 54, 12–14.
Wilkes, G.L., Sohn, S., Tamami, B., 2004. Preparations of noniso-
cyanate polyurethane materials from epoxidized soybean oils
and related epoxidized vegetable oils, Application: WO 2004-
US4823, 30 pp.
Young, C.S., Dolan, J.W., 2004. Success with evaporative light-
scattering detection. Part 2. Tips and techniques: LCGC North
America 22 (244), 246–250.
Hang, X., Yang, H., 1999. Model for a cascade continuous epoxida-
tion process. J. Am. Oil Chem. Soc. 76, 89–92.
Heron, S., Dreux, M., Tchapla, A., 2004. Post-column addition as a
method of controlling triacylglycerol response coefficient of an