Comparison of Two Types of Gas Chromatography on Fatty Acids
in Olive Oil
Adam Merrick, Berrien County Mathematics and Science Center, Berrien Springs,
There are new techniques for gas chromatography, one of them being
comprehensive 2-dimensional GC. This technology was applied to the study of fatty acid
profiles of olive oil. The samples were taken from extra virgin olive oil. Samples were
heated to find the difference in their fatty acid profiles before and after heating. The
separation of peaks that the comprehensive 2-dimensional GC provided was able to bring
forth additional information that would have gone unnoticed with a 1-dimensional GC.
Olive oils contain fats. Fats are triglycerides which have the chemical structure as
shown in Figure 1.
Figure 1. Chemical structure of a triglyceride.
Triglycerides contain two components, glycerol and fatty acids. The chemical structure
of glycerol is shown in Figure 2.
Figure 2. Chemical structure of glycerol.
Fatty acids are carboxylic acids with a hydrocarbon chain varying in length
between 12 and 24 carbons for olive oil. See Figure 3.
Figure 3. Chemical structure of fatty acid.
These hydrocarbon chains may have 0, 1 or more double bonds along this chain.
Chains with no double bonds are called saturated. Chains with at least one bond are
unsaturated. Specifically those with one double bond are called monounsaturated. Those
with two or more are called polyunsaturated. In the case that a double bond occurs on the
chain, it can orient itself in two ways. One way is called trans. The trans configuration
retains the zigzag shape of the chain. Cis is the other possibility, causing the chain to
bend or curve making it lose its straight structure. Figure 4 shows examples of a cis and
trans fatty acid.
Figure 4. Chemical structures of cis and trans fatty acids.
One way of naming fatty acids is by calling it “C “ followed by the number of
carbons in the chain, followed by a colon and the number of double bonds in the chain.
An example using figure 4 would be C18:1 cis, or C18:1 trans.
Olive oil has the following fatty acids present primarily within triglycerides:
lauric acid (C12:0) not present in discernible amounts, myristic acid (C14:0) <0.1%,
palmitic acid (C16:0) 7.5-20%, palmitoleic acid (C16:1) 0.3-3.5%, heptadecanoica acid
(C17:0) <0.5%, heptadecenoic acid (C17:0) <0.6%, stearic acid (C18:0) 0.5-5%, oleic
acid (18:1) 55-83%, linoleic (C18:2) 3.5-21%, linolenic acid (18:3) <1.5%, arachidic acid
(C20:0) <0.8%, behenic acid (C22:0) <0.3%, erucic acid (22:1) not present in discernible
amounts, lignoceric acid (C24:0) <1% Codex Alimentarius (1).
Olive oils are important to understand in their structure because they contain
many important and particular fatty acid structures. Olive oils are an important source of
monounsaturated fatty acids which have a healthy impact on one's plasma cholesterol
level according to Buckley and Howe (2). Perez-Jimenez et al. (3) described how olive
oil has been shown to be linked to better quality of life, longevity, and a lower incidence
of cardiovascular disease, cancer, and cognitive degeneration. Omega-3 fatty acids are
also known to be contained within olive oil and omega-3 fatty acids have been shown to
reduce fat levels in the body when already obese, thus giving olive oil that ability. The
omega-3 fatty acids change the gene expression of different parts of the body promoting
fat oxidation and energy usage instead of fat deposition.
Fatty acids that are heated in the presence of oxygen undergo oxidation.
Oxidation in cases has been shown to lower the amount of unsaturated fatty acids in oils.
Cossignani et al. (4) studied changes in olive oil fatty acid composition due to microwave
heating. GC analysis showed losses of unsaturated fatty acids. Caponio et al. (5) stated
that when vegetable oils are heated their fatty acid nutritional values worsen because
there is a lower content of unsaturated fats, which is shown to have healthy benefits.
In order to understand the biology of fats and oils, it is necessary to determine
their fatty acid composition. The typical analysis of fatty acids is done by analyzing the
methyl esters of the fatty acids by gas chromatography (GC). The fatty acids in the oil
must be converted (derivatized) to fatty acid methyl esters (FAME) in order for them to
be volatile enough to be analyzed by GC. When the fatty acids are derivatized they go
through three chemical reactions. The first of which is hydrolysis. To hydrolyze the
triglyceride you must add potassium hydroxide to it to separate the glycerol from the fatty
acid. This is done by a hydrogen being bonded in the place of the fatty acid on the
glycerol and a negative oxygen and a positive potassium being bonded to the fatty acid.
The negative oxygen and positive potassium are what’s called a potassium salt. The
second reaction is two reactions done together, neutralization and esterification. The fatty
acid potassium salts mix with sulfuric acid to give you potassium sulfate (K2SO4), a
white precipitate, and the fatty acid chain with hydrogen on the end instead of potassium.
The sulfuric acid then catalyzes the reaction of the fatty acid with methanol. The
methanol will make the hydroxyl (OH) on the end of the chain change into a methoxy
(OCH3) and give you water as well. This reaction is reversible which means that excess
methanol is needed to force the reaction to the right to the ester. At the end of the
reactions the FAMEs are extracted into hexane. The FAMEs prefer to partition into
hexane rather than the water/methanol mixture used for these reactions. Since hexane
does not mix with the water/methanol mixture, the final product of fatty acid methyl
esters or FAMEs in the hexane layer is removed from the two-layer mixture and analyzed
by gas chromatography (GC). Figure 5 shows these reactions.
Figure 5. Chemical reactions involved in derivatization of fatty acid to fatty acid methyl
The specific derivatization procedure used in these experiments followed the
procedure developed by O’Fallon et al. (6).
There is a standard GC method by the AOAC, method number 996.06 titled Fat
(Total, Saturated, and Unsaturated) in Foods, Hydrolytic Extraction Gas
Chromatographic Method (7). This method is a conventional one-dimensional GC
method. Recently, a method using comprehensive two-dimensional GC (GCxGC) has
been compared to the one-dimensional method and found to provide additional
information on the fatty acid composition, Tranchida et al. (8).
A comprehensive 2-dimensional Gas Chromatograph has one column with a
stationary phase, a viscous liquid coating the inside of the column, a modulator, a second
column with a different stationary phase than the first column, and the detector. What a
comprehensive 2-dimensional gas chromatograph does differently from a 1-dimensional
gas chromatograph, is that when analytes have gone through the 1st column those
separated remain so, but those that have not been separated can be separated in the 2nd
column with a different stationary phase. The purpose of the modulator is to preserve
separations made in the 1st column. Analytes separated may merge back together in a
different stationary phase, but with the modulator, analytes are slowed to a stop and
remobilized into the 2nd column at different times to ensure that analytes do not merge.
Figure 6 shows a simple diagram of a comprehensive 2-dimensional gas chromatograph.
Adahchour, M. (9) describes this technique in further detail.
Figure 6. Diagram of GCxGC instrument.
Tranchida et al. (8) showed how GCxGC-FID can provide additional information
(better separation) of FAMEs in olive oil and hazel nut oil. In these experiments we
compared GC and GCxGC for the analysis of FAMEs in olive oil that has been heated
and held at a cooking temperature of 350˚F and heated to the smoke point. The purpose
of this experiment is to find if GCxGC gives better results as compared to 1-dimensional
gas chromatography when comparing the effects of heat on olive oil. The hypothesis is
that the GCxGC method will give more information pertaining to olive oil FAMEs. This
is from the fact that GCxGC has more resolving power, i.e. being able to separate more
peaks, than GC.
Materials and Methods
Reagents. The following chemicals were used: ACS reagent grade hexane
(Sigma-Aldrich), ACS reagent grade methanol (Sigma-Aldrich), potassium hydroxide
pellets ACS certified (Fisher Scientific) diluted to 10N, concentrated sulfuric acid (Fisher
Scientific) diluted to 24N.
Standards. Standards used were: Food/Industry FAME mix (Restek product
number 35077). A 37 component FAME mix, 30mg/mL concentration.
Test Sample: Bertolli extra virgin olive oil.
The oil was heated to 350˚F in a metal pan as if it was used in cooking. Starting
from room temperature it took 10 minutes to reach 350˚F. Samples of the oil were taken
at 0, 15, 30, 45, and 60 minutes of 350˚F. The oil was continued to be heated to its smoke
point. Samples of the oil were taken at 0 and 5 minutes at the smoke point temperature.
The method of derivatization of the oils was as stated by O’Fallon (6). However
some changes were made to the method. The first is that there was no grinding or cutting
of samples, as the only samples were oils (40 µL). Another change is that instead of
being incubated in water, the sample tubes were put into the GC oven for the same
amount of time and temperature. The final change is that tubes were hand shaken instead
of being vortex mixed after adding hexane.
These are the conditions set for the 1-dimensional method: column flow of 2
mL/min of helium carrier gas, inlet temperature of 250˚C, split injection ratio of 200:1,
GC oven at 100˚C, 4 minute hold, 3˚C/min to 240˚C, 15 minute hold, secondary oven at
150˚C, 4 minute hold, 3˚C /min to 290˚C, 15 minute hold. There was no modulation
period. The flame ionization detector (FID) had a temperature of 250˚C, hydrogen flow
of 40 mL/min, air flow of 450 mL/min and helium makeup gas flow plus column flow of
50 mL/min. The data acquisition rate was 5 Hz. The conditions for the GCxGC are the
same except for the addition of the modulation period and the data acquisition rate (200
Hz). The conditions for the modulator were a 2 second modulation period and 0.6 second
hot pulse. A time-of-flight mass spectrometer was also used in place of the FID.
There were three columns installed in the GC. Column 1 was 100 m x 0.25 mm x
0.2 µm RT-2560 (Restek, cat#13199). Column number 2 was 0.5 m x 0.25 mm x 0.25
µm Rxi-lms, (Restek, cat#13320). The transfer line used was uncoated and was 0.35 m x
0.32 mm ID (J&W product#160-2325).
Results and Discussion
1D GC found higher amounts of 3 compounds due to heating, one peak however
could have been identified incorrectly as it had the same retention time as a fatty acid in
the standard. GCxGC was able to get these 3 new peaks as well, but with the addition of
the second dimension the previously spoken peak was able to be distinguished as
different from the fatty acid in the standard. There were none of the 3 peaks in the
unheated sample, but these peaks gradually increased with increasing time at 350 ˚C and
an increase to smoke point. All of the peaks were unable to be identified as fatty acids.
Figure 7 shows all the FAMEs in the standard using 1-dimensional GC. Figure 8
shows these same peaks but in a contour plot using GCxGC. Peaks are the same with the
first retention time but with GCxGC peaks can be separated visually by a second
retention time on the y-axis. In figure 9 the 3 smaller peaks within the 2680-2700 range
are the new compounds found. Figure 9 has a line in lime that is the unheated olive oil
sample with hardly any peaks apparent in the results. As the oil was heated the samples
show peaks appearing in greater heights the longer heat was applied. The large peak
going off the chart is of the FAME standard component #26. Since the peak of the
standard overshadows the new peak, if the first new peak was to be identified by using
the FAME standard in 1-dimensional GC, it would be identified as compound #26.
Figure 10 shows the 5 minute smoke point results in 2-dimensional GC, the
numbers given to peaks say that those peaks are compounds that correspond to the FAME
standard. The first of the 3 new peaks still has the same first retention time as compound
#26, however their second retention time is different, meaning they are different
compounds. Figure 11 shows the FAME standard using GCxGC and the empty box
shows where the new peaks would have been if included in those results.
Figure 12 is of the same separation of the 5 minute smoke point sample done in
the above results, except done with a time-of-flight mass spectrometer as the detector.
Figure 13 is of the mass spectrum of the first unknown peak in the 5 minute smoke point
sample, and figure 14 is the mass spectrum of the compound with the same retention
time: #26 from the FAME standard. In both figures are shown the measured spectrum
(peak true) and the best library match spectrum. Note that the standard’s similarity
number is 945 out of 1000, 1000 being a perfect match. The similarity number for the
unknown peak is 697 out of 1000. With a match of 697 the unknown is most likely not
the same as the library match. Looking at both the unknown’s spectrum and the
standard’s spectrum it is apparent that they are different compounds.
The hypothesis was that comprehensive 2-dimensional GC’s increased separation
capabilities gave greater amounts of information for the analysis of olive oil FAMEs.
Based on the results of our experiment it was confirmed. One of three peaks noted due to
the heating of olive oil would have been identified incorrectly in the 1-dimensional GC
analysis, but in the 2-dimesional GC analysis this peak and the other two were separated
from all the FAMEs in the standard mix.
Figure 7. 1-Dimensional chromatogram of the FAME standard. Starting from the 2nd
peak after 500 use table 1.
Figure 8. FAME standard using GCxGC, compounds are still in the same order as the 1-
D GC FAME standard.
Figure 9. Peaks of 1-dimensional results for new peaks
Figure 10. 2-dimensional results zoomed in around new peaks for smoke point sample of
Figure 11. 2-dimesional results for FAME standard zoomed in around where new peaks
Table 1. List of compounds in the FAME standard.
Figure 12. 5 minute smoke point sample shown using GCxGC-TOFMS focusing on only
Figure 13. Mass spectrum of first unknown peak.
Figure 14. Mass spectrum of compound #26 in the FAME standard.
1. Codex Standard for Olive Oil, Virgin, and Refined, and for Refined Olive-Pomace Oil.
Olive Oil Codex Stan 33 (1989). Codex Alimentarius, 8, 2001.
2. Buckley, J. D., & Howe, P. (2009). Anti-Obesity Effects of Long-Chain Omega-3
Polyunsaturated Fatty Acids. International Association for the Study of Obesity, 10, 648-
3. Perez-Jimenez, F., Ruano, J., Perez-Martinez, P., Lopez-Segura, F., & Lopez-Miranda,
J. (2007, March 20). The Influence of Olive Oil on Human Health: Not a Question of Fat
Alone. Mol. Nutr. Food Res,, 51, 1199-1208.
4. Cossignani, L., Simonetti, M. S., Neri, A., & Damiani, P. (1998). Changes in Olive Oil
Composition Due to Microwave Heating. JAOCS, 75(8), 931-937.
5. Caponio, F., Pasqualone, A., Gomes, T. (2003). Changes in the Fatty Acid
Composition of Vegetable Oils in Model Doughs Submitted to Conventional or
Microwave heating. International Journal of Food Science and Technology, 38, 481-486.
6. O'Fallon, J. V., Bushboom, J. R., Nelson, M. L., & Gaskins, C. T. (2007, February 12).
A Direct Method for Fatty Acid Methyl Ester Synthesis: Application to Wet Meat
Tissues, Oils, and Feedstuffs. Journal of Animal Science, 85, 1511-1521.
7. AOAC Official Method 996.06, Fat (Total, Saturated, and Unsaturated) in Foods,
Hydrolytic Extraction Gas Chromatographic Method, First Action 1996, Revised 2001
8. Tranchida, P. Q., Giannnino, A., Mondello, M., Sciarrone, D., Dugo, P., Dugo, G., &
Mondello, L. (2008). Elucidation of Fatty Acid Profiles in Vegetable Oils Exploiting
Group-Type Patterning and Enhanced Sensitivity of Comprehensive Two-Dimensional
Gas Chromatography. J. Sep. Sci., 31, 1797-1802.
9. Adahchour, M. (2009). Comprehensive Two Dimensional Gas Chromatography (Vol.
55, ). In L. Ramos (Ed.), Comprehensive Two Dimensional Gas Chromatography (Vol.
55, , pp. 3-44). Elsevier.