Improvement in the gas chromatographic resolution of petroselinate from oleate

Article abstract of DOI:10.1007/s11746-006-1222-y

In the extraction of oils from seeds of the genus Coriandrum, GC separations of petroselinate from oleate often gave poor resolution of these two isomers. Oleic and petroselinic acids were esterified with a series of alcohols (methanol, ethanol, 1 propanol, 2-propanol, 2-methyl-1 -propanol, 1 -butanol, 3-methyl1-butanol, and 2-ethyl-1-hexanol). GC resolution of the A6 from the Δ9 and Δ11 octadecenoates was examined for all ester derivatives on a polar phase column. The Δ6 and Δ9 isomers were unresolved as methyl esters; however, the 2-ethyl-1-hexyl esters gave baseline separation of all three isomers under temperature programming conditions. When isothermal conditions were optimized for each ester, separation of these isomers was possible with good resolution values (>89%) for all the alcohols except methanol, which had a partial resolution of 51%. The rates of esterification of all the alcohols were determined for reactions with both oleic acid and triolein using potassium hydroxide as the esterification catalyst. Methanol gave the largest rate constant in both acid and oil esterification reactions with a rate constant 10-fold better than all of the other alcohols. Based on rates of reaction, resolution of petroselinate from oleate, and removal of residual alcohol, the ethyl ester derivative appears to be the best choice for seed oils containing petroselinic acid. Copyright

Full text of DOI:10.1007/s11746-006-1222-y

Improvement in the Gas Chromatographic Resolution
of Petroselinate from Oleate
Terry A. Isbell*, Lindsay A. Green, Stephanie S. DeKeyser,
Linda K. Manthey, James A. Kenar, and Steven C. Cermak
USDA, ARS, National Center for Agricultural Utilization Research, Peoria, Illinois 61604
ABSTRACT: In the extraction of oils from seeds of the genus Co- niques have been devised. Kleiman (3) used a combination of
riandrum, GC separations of petroselinate from oleate often gave silver ion chromatography to isolate the monoene fraction of
poor resolution of these two isomers. Oleic and petroselinic acids
were esterified with a series of alcohols (methanol, ethanol, 1-
propanol, 2-propanol, 2-methyl-1-propanol, 1-butanol, 3-methyl-
Umbelliflorae seed oils followed by ozonolysis and GC quan-
tification of the cleavage products. Santinelli (4) epoxidized the
olefins and then converted the epoxides to chlorohydrins via a
hydrochloric acid-mediated ring opening. The resulting chloro-
hydrins were converted to their trimethylsilyl ethers, which
were partially resolved by GC. Thies (5) treated the FA with
dimethyldisulfide to make the bis-methylthio derivatives,
which gave only partial separation by GC.
1-butanol, and 2-ethyl-1-hexanol). GC resolution of the ∆6 from
the ∆9 and ∆11 octadecenoates was examined for all ester deriv-
atives on a polar phase column. The ∆6 and ∆9 isomers were un-
resolved as methyl esters; however, the 2-ethyl-1-hexyl esters
gave baseline separation of all three isomers under temperature
programming conditions. When isothermal conditions were opti-
mized for each ester, separation of these isomers was possible
Good separation of the ∆6 and ∆9 isomers has been reported
with good resolution values (>89%) for all the alcohols except using phenylethyl (6), isopropyl (7), and butyl (8–10) esters.
methanol, which had a partial resolution of 51%. The rates of es- The phenylethyl ester gives near baseline resolution on a 15-m
terification of all the alcohols were determined for reactions with
both oleic acid and triolein using potassium hydroxide as the es-
terification catalyst. Methanol gave the largest rate constant in
both acid and oil esterification reactions with a rate constant 10-
fold better than all of the other alcohols. Based on rates of reac-
tion, resolution of petroselinate from oleate, and removal of resid-
ual alcohol, the ethyl ester derivative appears to be the best
choice for seed oils containing petroselinic acid.
polar phase (DB-23) column when the esterification reaction
conditions are mild using acid or base catalysis. The authors
reported that purification of the ester by a silica-gel TLC plate
separation was necessary to remove residual phenylethanol.
The isopropyl esters also provided near baseline separation, but
derivatization was slow due to the secondary nature of the al-
cohol. The best resolution reported appeared to be when butyl
esters were used but the authors report (8,9) that resolution was
lost after several injections due to fouling of the precolumn
Paper no. J11276 in JAOCS 83, 429–434 (May 2006).
KEY WORDS: Branched alcohols, coriander, esters, gas chro- when butanol was used as the derivatization solvent. This prob-
matography, oleic, petroselinic, rates.
lem was overcome using n-hexane as the solvent in the esterifi-
cation step.
Based on these literature reports, one might conclude that
Petroselinic acid is present in a number of seed oils but is most the driving force for good resolution of the ∆6 and ∆9 octade-
commonly found in the Umbelliferae family. One interest in cenoates is dependent on the ability to differentiate small po-
this acid is that it could serve as a useful raw material for pro- larity differences between these two isomers. Accordingly,
ducing adipic and lauric acids via oxidative cleavage of the ∆6 masking of the polar carboxyl group with larger alkyl ester
double bond in petroselinic acid. Adipic acid is commonly used groups limits the interaction of carboxyl oxygens with the polar
as a monomer for nylon synthesis, and lauric acid is used in de- phase of the capillary column. A detailed study of the effect of
tergent applications. Analysis of petroselinic acid in natural oils alcohol chain length and branching on ester resolution for pet-
by GC is typically confounded by the presence of oleic acid. roselinate from oleate, however, has not been reported. Christie
When methyl esters of petroselinic and oleic acids are synthe- (10) refers to a good resolution of the ∆6 and ∆9 octade-
sized, the GC separation of these two isomers is incomplete cenoates and cites unpublished data for a series of short-chain
(
1,2). Separation of petroselinate from oleate was reported to alcohols. In addition, the esterification rate of these FA with
have only a 0.04 equivalent chain length (ECL) unit resolution, this alcohol series will be important since rapid derivatization
which made quantification of the individual isomers difficult. is often a requirement for analyzing large numbers of
In an effort to overcome these quantification problems, several germplasm samples associated with breeding programs.
tedious chromatographic separation or derivatization tech-
*
1
To whom correspondence should be addressed at USDA-ARS-NCAUR, EXPERIMENTAL PROCEDURES
815 N. University St., Peoria, IL 61604. E-mail: isbellta@ncaur.usda.gov
Materials. Coriander oil was a gift from International Flora
Technologies Ltd. (Gilbert, AZ). Hexanes, acetone, methanol,
Current address of third author: University of Wisconsin, 5234 Rennebohm
Hall, 777 Highland Ave., Madison, WI 53705, USA.
Copyright © 2006 by AOCS Press
429
JAOCS, Vol. 83, no. 5 (2006)

430
T.A. ISBELL ET AL.
1
3
2
-propanol, 2-propanol, 1-butanol, 2-methyl-1-propanol, and
-methyl-1-butanol were from Sigma-Aldrich (St. Louis, MO).
-Ethyl-1-hexanol was from Acros (Fairlawn, NJ). Ethanol
was obtained from Aaper (Shelbyville, KY). FAME standards
were from Alltech Associates (Deerfield, IL).
GC. GC was performed with a Hewlett-Packard 5890 Series
II gas chromatograph (Palo Alto, CA), equipped with a FID and
an autosampler/injector. Analyses were conducted on an SP-
2380 column (0.25 mm × 30 m × 0.2 µm film thickness; Supelco,
f
Bellefonte, PA). GC analyses were conducted as follows: col-
umn flow 1.0 mL/min with helium head pressure of 16.9 psi (116
kPa) using electronic pressure control set to maintain constant
flow; split ratio 200:1. Temperature ramp analysis: programmed
ramp 150 to 265°C at 10°C/min with injector and detector tem-
peratures set at 250°C. Under isothermal analysis conditions, all
GC parameters were the same for each alcohol series, except the
FIG. 1. Graphical depiction of calculation for resolution of unresolved
peaks, θ = f/g.
temperature was held at the values listed in Table 3 (presented calculations and was determined by the ECL for 18:0 = 18.0
shortly) for each ester to obtain a 10 min run. Saturated C –C and the ECL for 18:2 = 19.952, which were calculated as pre-
FAME provided standards for calculating ECL values, which viously described.
8
30
were used to make FAME assignments.
HPLC for esterification rate study. HPLC analyses were
Resolution calculation. The resolution for the separation of performed on a Thermo Separations P2000 binary gradient
∆
6 from the ∆9 octadecenoates and the separation of the ∆9 pump (San Jose, CA) with a Spectra System AS3000 autosam-
from the ∆11 octadecenoates was calculated using the method pler/injector from Thermo Separation Products (Fremont, CA)
of Kaiser (11). The following equation was used to calculate coupled to an ELSD from Alltech Associates (Deerfield, IL).
the resolution; θ = f/g where θ is the resolution as a value be- A Dynamax (250 mm × 4.6 mm, 60 Å, 8 µm) silica column
tween 0 and 1.0, which can be represented as a percentage by purchased from Rainin Instrument Co., a division of Varian
multiplying θ by 100. Figure 1 provides a graphical depiction (Walnut Creek, CA), was used to separate the mixtures. Com-
of the calculation. The term f represents the distance between ponents were eluted from the column using the following gra-
the valley at point v and point m on a line that connects the dient elution: 0–1 min, isocratic hexane/ethyl acetate 95:5; gra-
maxima of peaks 6 and 9. The term g represents the distance dient to 90:10 at 2 min then hold until 5 min followed by re-
from the baseline to point m on the same line connecting peaks verse gradient to 95:5 at 5.5 min then isocratic to 7.0 min. A
6
and 9. This method is particularly suited to adjacent peaks flow rate of 1 mL/min was used. The ELSD drift tube was set
whose heights differ greatly. The method calls for a graphical at 50°C with the nebulizer set at 20 psi (138 kPa) N to give a
2
determination of the point m. We calculated the equation for flow rate of 2.0 standard liters per minute (SLPM). Standard
this line to eliminate the graphical errors that would result from curves for both triolein and oleic acid were developed by serial
determining an accurate intersect point (m) on a line with this dilution of a 0.50 mg/mL standard to cover the range 0.50–0.03
steep of a slope.
mg/mL for a total of seven concentrations. Response curves for
Relative ramp rate calculation. The relative ramp rate for triolein and oleic acid were then developed under these analy-
each alcohol studied under the temperature program analysis sis conditions.
conditions was calculated from the change in oven temperature
General esterification reactions for GC resolution study. Es-
to elute the C18:0 peak compared with the C18:2 peak divided ters were made by dissolving 2–10 mg of coriander oil in 0.5
by the change in ECL as expressed in Equation 1:
mL of a solution of the appropriate alcohol that is 0.5 M in
potassium hydroxide in a 2-mL crimp-top vial. The vial was
[1] sealed with a crimp cap and placed in a heating block main-
tained at 60°C for 1 h. The vial was removed from the heating
relative ramp rate = ∆TC18:2 – C18:0/∆ECL
The temperature ramp program described previously had an block and, after cooling to room temperature, the cap was re-
initial temperature of 150°C with a temperature ramp rate of moved and 0.5 mL of 1.0 M sulfuric acid in the same alcohol
1
0°C/min. Therefore, the oven temperature for the retention was added to the vial. The vial was sealed with a new cap and
time of each isomer was determined by Equation 2:
placed back in the heating block at 60°C. After 15 min the vial
was removed and allowed to cool to room temperature; then the
[2] cap was removed. The vial content was transferred to a 2-dram
vial using 2 mL of hexane; 1.0 mL of saturated brine was added
150°C + (10°C × retention time) = Tpeak
Stearate (18:0) and linoleate (18:2) were present in all the sam- to the vial, and the solution was then mixed thoroughly. After
ples and were used as an internal standard to calculate ECL val- the water phase separated, 1.0 mL of the upper hexane layer was
ues across the octadecenoate isomer region. Therefore, the drawn off with a Pasteur pipette and placed in a 2.0-mL target
change in ECL was the same value (∆ECL = 1.952) for all the vial and analyzed by the GC method previously described.
JAOCS, Vol. 83, no. 5 (2006)

GC OF PETROSELINATES
431
TABLE 1
concentration, type of catalyst, reaction temperature, and alco-
hol concentration, making direct comparisons difficult. Be-
cause of these discrepancies, a series of straight and branched-
chain alkyl alcohols was used to examine the esterification rate
of oleic acid and triolein under pseudo first-order kinetics.
These data are presented in Tables 1 and 2.
In the reactions of oleic acid with this series of alcohols
(Table 1), the rate constant ranged across two orders of magni-
tude; methanol provided the fastest rate and 2-propanol the
slowest. The rate for methanol was significantly faster (nearly
Rate of Esterification of Oleic Acid
Alcohol
Rate constant (k)
Relative rate
−1
−2
−2
−3
−3
−2
−2
−3
Methanol
Ethanol
1.78 × 10
2.37 × 10
1.64 × 10
8.06 × 10
9.41 × 10
1.59 × 10
1.47 × 10
9.37 × 10
22.1
2.9
2.0
1.0
1.2
2.0
1.8
1.1
1
2
2
1
3
2
-Propanol
-Propanol
-Methyl-1-propanol
-Butanol
-Methyl-1-butanol
-Ethyl-1-hexanol
10-fold) than for any of the other alcohols studied. As expected,
as the alkyl chain length increased within the straight-chain pri-
mary alcohols, the rate of reaction decreased owing to de-
creased nucleophilicity of the alcohol (methanol >>> ethanol >
1-propanol > 1-butanol). Furthermore, the addition of branch-
ing at the β-carbon of the primary alcohol increased the steric
bulk of the alcohol, further diminishing the rate of esterifica-
tion, as can be seen by comparison of 1-propanol with 2-
methyl-1-propanol. The only secondary alcohol studied, 2-
propanol, gave the slowest rate, as expected.
For esterifications with triolein and this series of alcohols
(Table 2), the rate constant ranged across two orders of magni-
tude, where methanol again gave the fastest rate and 2-propanol
the slowest. The rate of methyl ester formation compared with
TABLE 2
Rate of Esterification of Triolein
Alcohol
Rate constant (k)
Relative rate
−2
−3
−3
−4
−3
−3
−3
−3
Methanol
Ethanol
1.23 × 10
3.29 × 10
3.02 × 10
1.75 × 10
3.27 × 10
4.34 × 10
3.97 × 10
3.82 × 10
70.3
18.8
17.2
1.0
18.7
24.8
22.7
21.8
1
2
2
1
3
2
-Propanol
-Propanol
-Methyl-1-propanol
-Butanol
-Methyl-1-butanol
-Ethyl-1-hexanol
Esterification rate studies. Pseudo first-order reaction con- the rest of this series was not as dramatic as those reactions
ditions were made by dissolving oleic acid (0.60 g, 2.13 mmol) with oleic acid, with only an approximately threefold faster
or triolein (0.50 g, 0.56 mmol) in a solution of 1.0 M sulfuric rate. The leveling of the rate constant across the entire series
acid in alcohol (100 mL) and placed in a constant-temperature was most likely a result of increased solubility of the triolein in
reactor connected to a constant-temperature circulating bath the longer-chain alcohols (1-butanol > ethanol). In addition,
maintained at 30°C. All reactions were mixed via a magnetic branching at the β-carbon of the alcohol appears to have little
stir bar. Consumption of oleic acid or triolein was followed by impact on the rate of these primary alcohols (1-propanol vs. 2-
removing 1-mL aliquots at designated time intervals, diluting methyl-1-propanol). However, esterification with the sec-
with 1 mL of hexane, and then injecting on the high-perfor- ondary alcohol, 2-propanol, is quite slow. Lastly, all the triolein
mance liquid chromatograph and analyzing as previously de- esterifications are nearly an order of magnitude slower than
scribed. All reactions were run in duplicate, and initial rates their corresponding reaction with oleic acid.
were determined by plotting the log of the oleic or triolein con-
GC resolution of octadecenoate positional isomers. The al-
centration vs. time. A linear regression of the first 4–8 data cohols listed in Tables 1 and 2 were used to esterify coriander
points of each plot provided a slope, which represents the ini- oil and examine the GC resolution of ∆6 (petroselinic), ∆9
tial rate of esterification. The average of the two esterification (oleic), and ∆11 (vaccenic) esters. The resulting coriander esters
rates for each alcohol studied (Tables 1, 2) was used to calcu- were analyzed by GC using both temperature programming and
late the rate constant k (k = −2.303 × slope) for all reactions.
isothermal analysis conditions on a polar (90% bis-
The esterification rate study was conducted at 30°C (rather cyanopropyl/10% cyanopropylphenyl siloxane) 30 m capillary
than 60°C) so that the rate of reaction was slow enough for a column. A temperature-programmed ramp that we typically use
sufficient number of data points to be collected before the reac- for screening seed oils with FA profiles having ECL values that
tion reached its half-life. In practice, the esterification reaction vary from C6 to C30 was used for the initial evaluation. Using
would typically be run at the highest temperature possible so this temperature program, separation of the petroselinate from
that the reaction could reach completion in minimal time with the oleate and vaccenate esters was conducted, and these aligned
no deleterious side product formation. We found 60°C to be a chromatograms are shown in Figure 2. Good separation of the
suitable temperature to meet this criterion.
isomers is observed as the chain length or the branching of the
ester increased. Under these analysis conditions, the 2-ethyl-1-
hexyl ester provided the highest resolution value for the separa-
tion of the ∆6 from the ∆9 isomer (Table 3, Resol. = 0.957 ∆6-
RESULTS AND DISCUSSION
Esterification rate studies. A number of rate studies for esteri- ∆9 column) whereas the methyl ester provided no resolution of
fication of oleic acid with several of the alcohols used in this these two isomers. These data suggested that sterically masking
study have been reported (12–16). Unfortunately, reaction con- the polarity of the carboxyl group, so the main interaction would
ditions between these studies varied widely in terms of catalyst be between the small polarity differences of the olefins and the
JAOCS, Vol. 83, no. 5 (2006)

432
T.A. ISBELL ET AL.
FIG. 2. Stack of GC traces of coriander esters from an SP-2380 column (0.25 mm × 30 m × 0.2 µm film thickness),
temperature ramp 150 to 265°C at 10°C/min. Chromatograms aligned based on the ∆6 octadecenoate.
polar liquid phase of the column, caused the observed separa- derived from the individual alcohols, and these chromato-
tions. However, the longer-chain esters have longer retention graphic separations are shown in Figure 3. To optimize the
times under this temperature ramp program, providing increased isothermal runs, conditions were found such that all the esters
time for partitioning with the column liquid phase, thereby en- would have retention times between 5 and 9 min, and good res-
hancing resolution. As a result, Table 3 shows that the shorter- olution of the isomers was observed. Under these conditions, all
chain esters have higher relative ramp rates per ECL (°C/ECL of the esters except methyl gave a ∆6, ∆9, and ∆11 separation
column of Table 3) over the 18:1 retention time range being that would be suitable for quantitative purposes. The tempera-
studied, thus making comparisons between the resolution data ture, retention time, and resolution data for these esters are
of these 18:1 isomers under temperature ramp conditions tenu- shown in the bottom half of Table 3. In the series ethyl through
ous. To ascertain the effects of alcohol chain length and branch- 2-ethyl-1-hexyl, the resolution indices are all >89% resolved for
ing on the resolution of the 18:1 isomers, the ramp rate per ECL the ∆6-∆9 octadecenoate peaks, and even the methyl ester is
should be equal in the region of interest. Although isothermal now 51% resolved. The 2-ethyl-1-hexyl ester is completely re-
conditions for the analysis of a broad range of FA chain lengths solved under the isothermal conditions with baseline separation.
typically give poor results, use of an isothermal run over a nar- Of particular note, the branched esters all tended to have lead-
row retention time span containing the desired FA, particularly ing peaks. Leading peaks can be caused by a number of param-
with the same chain length, can provide good resolution. Thies eters including column overload, poor injection techniques,
(8) has reported the use of isothermal conditions for the separa- poor solubility of the analyte in the stationary phase, and sam-
tion of the butyl ester of oleic acid isomers at 220°C with a total ple condensation in the system caused by too low of a column
run time of 13 min. Thies saw good resolution of the butyl es- temperature. Because the straight-chain esters gave good peak
ters under these conditions, but baseline separation was not ob- symmetry even though they were prepared to have similar con-
served. We optimized the isothermal runs for each set of esters centrations and injected by the same auto-injector as the
JAOCS, Vol. 83, no. 5 (2006)

GC OF PETROSELINATES
433
TABLE 3
GC Resolution of Esters
Temperature Relative ramp ratea
Ret. time ∆6
Ret. time ∆9
Ret. time ∆11
Ester
(°C)
(°C/ECL)
(min)
(min)
(min)
Resol. ∆6-∆9
Resol. ∆9-∆11
Methyl
Ethyl
Ramp
Ramp
Ramp
Ramp
Ramp
Ramp
Ramp
Ramp
165
170
175
165
190
2.9
2.8
2.8
2.5
2.6
2.7
2.6
2.3
0
0
0
0
0
4.710
4.793
5.114
4.679
5.220
5.449
5.609
6.426
8.728
7.756
7.875
8.838
5.121
6.784
5.291
6.825
4.710
4.845
5.140
4.736
5.250
5.479
5.638
6.464
8.791
7.848
7.991
8.956
5.175
6.892
5.360
6.956
4.770
4.932
5.174
4.782
5.276
5.513
5.670
6.497
8.963
8.009
8.163
9.163
5.255
7.043
5.447
7.105
0.000
0.745
0.616
0.855
0.845
0.910
0.819
0.957
0.512
0.894
0.955
0.935
0.922
0.977
0.967
0.992
0.987
0.945
0.970
0.997
0.969
0.968
0.983
0.989
0.985
0.922
0.998
0.983
0.949
0.997
0.988
0.992
1
2
2
1
3
2
-Propyl
-Propyl
-Methyl-1-propyl
-Butyl
-Methyl-1-butyl
-Ethyl-1-hexyl
Methyl
Ethyl
1
2
2
1
3
2
-Propyl
-Propyl
-Methyl-1-propyl
-Butyl
-Methyl-1-butyl
-Ethyl-1-hexyl
185
195
200
0
0
0
aRelative ramp rate (°C/ECL) was calculated from the change in oven temperature to elute the C18:0 peak compared with the C18:2 peak divided by the
change in ECL, which is 1.952 for all the entries.
FIG. 3. Stack of GC traces of coriander esters from an SP-2380 column (0.25 mm × 30 m × 0.2 µm film thickness).
Isothermal temperature programs with temperature are shown on trace. Chromatograms aligned based on the ∆6
octadecenoate.
JAOCS, Vol. 83, no. 5 (2006)

434
T.A. ISBELL ET AL.
branched esters, column overload and poor injection technique
can be readily eliminated as possible causes of the leading
peaks. However, it is possible that the branched esters exhibit
lower solubility in the stationary phase and therefore produce
leading peaks. To investigate the possibility that analyte was
condensing in the system due to low column temperature, we
increased the isothermal temperature above the corresponding
temperature of a straight-chain ester having the same number of
carbon atoms (2-methyl-1-propyl, 190°C vs. 1-butyl, 185°C).
This increase in temperature resulted in improved peak symme-
try. Unfortunately, the leading peak could not be sharpened for
the 2-ethyl-1-hexyl ester even with increased temperature. To
explore the possibility of improving peak resolution, lower
isothermal temperatures, which result in longer retention times,
were examined (reported in Table 3). However, decreasing the
temperature did not improve peak resolution and leading peaks
persisted.
Based on the rates of esterification and the resolution of iso-
mers, the 2-ethyl-1-hexyl ester would appear to be promising
for the quantitative analysis of oils containing a mixture of pet-
roselinic, oleic, and vaccenic acids. However, 2-ethyl-1-hexa-
nol was not appreciably soluble in water, and removal of ex-
cess alcohol in the esterification step by a water wash did not
occur in those alcohols larger than 1-propanol. As a result,
The Determination of the Petroselinic Acid Content of Corian-
der (Coriandrum sativum) Oil by Capillary Gas Chromatogra-
phy, Phyto-Chemical Analysis 3:250– 253 (1992).
. Kleiman, R., and G.F. Spencer, Search for New Industrial Oils:
XVI. Umbelliflorae-Seed Oils Rich in Petroselinic Acid, J. Am.
Oil Chem. Soc. 59:29–38 (1982).
4. Santinelli, F., and P. Damiani, A Simple and Rapid Method for
the Concurrent Determination of Petroselinic and Oleic Acids
in Oils, Ibid. 74:935–938 (1997).
. Thies, W., Determinations of the Petroselinic Acid Content in
Seeds of Coriandrum sativum by Gas Liquid Chromatography,
Fat Sci. Technol. 95:20–23 (1993).
3
5
6. Liu, L., and E.G. Hammond, Phenylethyl Esters of Fatty Acids
for the Analytical Resolution of Petroselinate and Oleate, J. Am.
Oil Chem. Soc. 72:749–751 (1995).
7
. Wolff, R.L., and F.F. Vandamme, Separation of Petroselinic
cis-6 18:1) and Oleic (cis-9 18:1) Acids by Gas–Liquid Chro-
(
matography of Their Isopropyl Esters, Ibid. 69:1228–1231
(1992).
8
. Thies, W., Determination of the Petroselinic Acid in Seeds of
Coriandrum sativum by Gas Liquid Chromatography as n-Butyl
Esters, Fat Sci. Technol. 97:411–413 (1995).
. Reiter, B., M. Lechner, and E. Lorbeer, Determination of Pet-
roselinic Acid in Umbelliferae Seed Oils by Automated Gas
Chromatography, J. High Resolut. Chromatogr. 21:133–136
9
(
1998).
1
0. Christie, W.W., Are Methyl Esters of Fatty Acids the Best
Choice for Gas Chromatographic Analysis, Lipid Technol.
7:64–66 (1995).
residual alcohol causes longer separation times during the ex- 11. Kaiser, R., Gas Phase Chromatography, Butterworth, London,
traction step of the esterification reaction due to emulsion for-
mation. In addition, injection of these residual alcohols onto
the gas chromatograph is known (8,9) to be detrimental to iso-
mer resolution owing to fouling of the injection liner. With
these practical considerations in mind, ethanol may be the al-
cohol of choice to provide a reasonable rate of esterification
and good isomer resolution along with rapid partitioning and
easy removal of residual alcohol from the reaction mixture.
1963, Vol. I, pp. 39–40.
1
1
1
2. Warabi, Y., D. Kusdiana, and S. Saka, Reactivity of Triglyc-
erides and Fatty Acids of Rapeseed Oil in Supercritical Alco-
hols, Bioresour. Technol. 91:283–287 (2004).
3. Markley, K.S., Esters and Esterification, in Fatty Acids, Their
Chemistry, Properties, Production, and Uses, Part 5, edited by
K.S. Markley, Interscience, New York, 1968, pp. 3459–3464.
4. Tartarelli, R., G. Stoppato, F. Morelli, A. Lucchesi and M.
Giorgini, Esterification of Oleic Acid with Butanol Catalyzed
by a Cation-Exchanged Resin, J. Catalysis 13:108–110 (1969).
5. Othmer, D., and R.S. Anada, Butyl Oleate from Butyl Alcohol
and Oleic Acid, J. Ind. Eng. Chem. 42:1912–1919 (1950).
6. Freedman, B., R.O. Butterfield and E.H. Pryde, Transesterifica-
tion Kinetics of Soybean Oil, J. Am. Oil Chem. Soc.
63:1375–1380 (1986).
1
1
REFERENCES
1
. Christie, W.W., Equivalent Chain-Lengths of Methyl Ester De-
rivatives of Fatty Acids on Gas Chromatography, J. Chro-
matogr. 447:305–314 (1988).
2
. Griffiths, D.W., G.W. Robertson, S. Millam, and A.C. Holmes,
[Received November 23, 2005; accepted February 20, 2006]
JAOCS, Vol. 83, no. 5 (2006)