Content uploaded by Frank David
Author content
All content in this area was uploaded by Frank David on Jan 09, 2015
Content may be subject to copyright.
Author
Frank David
Research Institute for Chromatography
Pres. Kennedypark 20, B-8500 Kortrijk
Belgium
Francis Scanlan
Quest International
Naarden, The Netherlands
Pat Sandra
Laboratory of Organic Chemistry
University of Gent
Krijgslaan 281, B-9000 Gent
Belgium
Michael Szelewski
Agilent Technologies, Inc.
2850 Centerville Road
Wilmington, DE 19808-1610
USA
Abstract
Two retention time locked methods for the analysis of
essential oil compounds are described. The first method is
a gas chromatography/flame ionization detector method
using a 50 m ××320 µm id ××1.05 µm HP-5 column. The
second method is a gas chromatography/mass spectrom-
etry method using a 30 m ××250 µm id ××0.25 µm
HP-5MS column. Retention times of approximately 400
essential compounds were measured using both methods,
Analysis of Essential Oil Compounds Using
Retention Time Locked Methods and
Retention Time Databases
Application
and two retention time databases were created. Flavor
compounds in food extracts or essential oil constituents
can be automatically searched based on retention times
and/or mass spectra. Finally, transformation of existing
retention index libraries into locked retention time
databases is discussed.
Introduction
Capillary gas chromatography (GC) has been for
many years the method of choice for the analysis
of essential oils [1]. The constituents of essential
oils are identified using a combination of different
GC techniques, including GC with flame ionization
detection (FID) and determination of retention
indices, GC with olfactometric detection (sniffing),
GC in combination with mass spectrometry
(GC/MS) and GC with element-selective detection
(flame photometric detection, nitrogen phospho-
rous detection, atomic emission detection, etc).
Although GC/MS is probably the most powerful
technique, and extended mass spectral libraries
are available, it does not allow complete identifica-
tion. Essential oils are complex mixtures of
monoterpenes, monoterpenoids, sesquiterpenes,
sesquiterpenoids, diterpenes, and diterpenoids. No
single capillary column can resolve all possible
compounds, and spectral data are not always con-
clusive because isomers give similar spectra. In
flavor and fragrance quality control, retention
indices are still frequently used as a complemen-
tary technique to GC/MS. Several libraries are
Food and Flavors
available with retention indices for many flavor
and fragrance compounds [2−4]. Retention indices
are less dependent on operational parameters than
absolute retention times, but they still depend sig-
nificantly on the column type (stationary phase
and supplier), on the temperature program, and to
a lesser extent, on the carrier gas velocity. There-
fore, it is sometimes difficult to reproduce pub-
lished retention indices in different laboratories.
Moreover, most companies in the flavor and fra-
grance industry are still using in-house methods
based on historical choices of columns and condi-
tions. Finally, the use of retention indices (with
n-alkanes as reference compounds) is not possible
with element-selective detectors. A peak detected
at a certain retention time using a sulphur selective
detector, for instance, might be difficult to locate in
an FID or MS chromatogram.
Recent developments in GC have led to the ability
of locking and matching retention times for a given
application [5-6]. Using retention time locking, it is
no longer necessary to calculate the retention
index, but the absolute retention time can be used
as an identification tool. Of course, retention times
are still dependent on operating conditions, but
small differences in carrier gas velocity and
column length are compensated by re-locking the
GC method by adjustment of the column head-
pressure. After re-locking, elution temperatures of
the solutes are also constant. Moreover, retention
time locking can also be used in combination with
different detectors, and exact scaling of GC/FID,
GC/sniffing, GC/MS, and GC/AED chromatograms
is possible [6]. Retention time locking and reten-
tion time databases are, therefore, excellent tools
in essential oil and in flavor QA/QC
analysis.
In this paper, two retention time locked methods
are presented. For each method, a retention time
database is available containing approximately 400
flavor compounds and essential oil constituents.
The first method is a GC/FID method. A long, thick
film column is used in combination with a slow
temperature program. These conditions are fre-
quently used in QA/QC analysis in the flavor and
essential oil industry because a high sample capac-
ity is combined with a high resolving power,
2
resulting in a detailed picture of the samples. The
second method is a GC/MS method. For this
method, a 30 m ×0.25 mm id ×0.25 µm HP-5MS
column was selected because this is the most fre-
quently used column in GC/MS analysis. While the
resolution and sample capacity are lower on this
column, the GC/MS analysis mainly focuses on
identification of analytes. For this method, a com-
bined retention time and mass spectral library
Screener Database is available.
Because a lot of retention data are already avail-
able as retention indices, it was also evaluated if
these data could be transferred into absolute
retention times that match with locked retention
times. It was shown that retention indices from
existing retention index libraries can be recalcu-
lated as absolute retention times that match with
experimental data.
Experimental
GC/FID analyses were performed on an Agilent
6980 gas chromatograph equipped with a
split/splitless inlet. Separation was done on a
50 m ×0.32 mm id ×1.05 µm HP-5 column (β=72)
(Agilent part number 19091J-215). The analytical
conditions are summarized in Table 1. Helium at
approximately 85 kPa (12.33 psi) constant pres-
sure was used as carrier gas. The inlet pressure
was adjusted to give a retention time of 70.000 min
for n-pentadecane. This is done by retention time
locking (RTL), using five runs at different pres-
sures (respectively 70, 80, 90, 100, and 110 kPa),
and plotting the retention time of n-pentadecane
as a function of the inlet pressure [5]. From this
curve, the inlet pressure can be calculated to
adjust the retention time of n-pentadecane to
exactly 70.000 min. The analytical conditions in
this GC/FID method are typical conditions used in
quality control of essential oils. The column choice
and the slow temperature program offer high reso-
lution and a detailed sample profiling. The column
also offers high sample capacity, which is also
important in essential oil profiling, because impor-
tant trace constituents can be present and elute
close to major constituents. On columns with
restricted sample capacity, overloading and, conse-
quently, peak leading is frequently observed for the
main constituents.
3
The second method is a GC/MS method. GC/MS
analyses were performed on an Agilent 6980 gas
chromatograph equipped with a split/splitless
inlet in combination with an Agilent 5973N MSD.
Separation was done on a 30 m ×0.25 mm id ×
0.25 µm HP-5MS column (β=250) (Agilent part
number 19091S-433). The analytical conditions are
summarized in Table 2. Helium at approximately
65 kPa (9.43 psi) constant pressure was used as
carrier gas. The inlet pressure was adjusted to give
a retention time of 27.500 min for n-pentadecane.
This is done by retention time locking, using five
runs at different pressures, and plotting the reten-
tion time of n-pentadecane as a function of the inlet
pressure. This is automatically done by starting the
“acquire RTL calibration runs” command in the
GC/MS instrument control. From this curve the inlet
pressure can be calculated to adjust the retention
time of n-pentadecane to exactly 27.500 min. These
analytical conditions can be used to screen essen-
tial oils using GC/MS. Essential oil constituents
can be identified based on the mass spectral data,
and on retention times, using a screener library.
The operational conditions are identical to the
conditions used by Adams [4]. Spectra and reten-
tion data published in this reference are also valid
for this method.
Test mixtures of flavor compounds and n-alkanes
were prepared from pure chemicals at 0.1% con-
centration in carbon tetrachloride or chloroform.
Essential oil mixtures are diluted in carbon tetra-
chloride or chloroform at a 5% level (50 mg/mL).
Results and Discussion
GC/FID Method
The described GC/FID method is used for quality
control of essential oil mixtures. The long, thick
film column results in high resolution and high
sample capacity. Traces of important compounds
can be detected besides the main constituents. A
typical separation obtained by this method
appears in Figure 1, showing the analysis of a
Spanish orange oil. The chromatogram shows a
detailed picture of the main compounds and of
minor constituents. This type of analysis, giving
both qualitative and quantitative information, is
used for quality control of essential oils. This
GC/FID method was locked to n-pentadecane
(t
R
= 70.000 min). Under these locked conditions,
n-decane elutes at 31.640 min and n-eicosane at
99.557 min. Using these conditions, a retention
time locked database was created containing
approximately 400 compounds that are important
in quality control of essential oil mixtures. Using
this database and the GC ChemStation RTL option,
peaks in the GC chromatogram can be identified
based on a retention time search in a given reten-
tion time window (for instance ±0.2 min). For the 10
main peaks of the Spanish orange oil, the results of
such a retention time search are given in Table 3. It
is clear, that in some cases, several compounds
elute in the 0.4-min window and further identifica-
tion is needed. However, this tool already allows an
excellent profiling of samples.
Column 50 m ×0.32 mm id ×1.05 µm HP-5 (β=72)
(Agilent part number 19091J-215)
Injection Split, split ratio 25:1, 250 °C, 1 µL injection
volume
Carrier Helium (approximately 85 kPa) (12.33 psi),
constant pressure
RTL The inlet pressure is adjusted to give a reten-
tion time of 70.000 min for n-pentadecane
Oven program 50 °C to 280 °C at 2 °C/min
(110 min analysis time)
Detection FID, 300 °C
Table 1. GC/FID Conditions
Column 30 m ×0.25 mm id ×0.25 µm HP-5MS (β=72)
(Agilent part number 19091S-433)
Injection Split, split ratio 25:1, 250 °C, 1 µL injection
volume
Carrier Helium (approximately 65 kPa) (9.43 psi),
constant pressure
RTL The inlet pressure is adjusted to give a reten-
tion time of 27.500 min for n-pentadecane
Oven program 60 °C to 240 °C at 3 °C/min (60 min analysis
time)
Detection MS in scan mode (40_400 amu);
solvent delay: 2 min; transfer line: 300 °C
Table 2. GC/MS Conditions
4
0102030405060 min
3
7
4
10
9
8
6
5
2
1
70
Figure 1. GC/FID chromatogram of Spanish orange oil. (Conditions: Table 1, peak identification: Table 3)
Table 3 Identification of main compounds in Spanish orange oil using a retention time database and a
combined mass spectral and retention time identification.
Peak GC/FID* GC/MS
number t
R
(min) t
R
identification t
R
(min) MS + t
R
identification
1 26.793 α-pinene 5.172 α-pinene
2 30.042 1-octen-3-ol 6.181 sabinene
3-(methylthio)-1-propanol
sabinene
3 30.539 hexanoic acid 6.282 β-pinene
β-pinene
6-methyl-5-hepten-2-one
4 31.053 2-octanone 6.658 myrcene
myrcene
furfuryl acetate
5 31.987 octanal 6.987 octanal
6 33.190 trans-2-hexenoic acid 7.267 ∆-3-carene
∆-3-carene
7 35.001 limonene 8.130 limonene
benzylalcohol
ocimene
5
Peak GC/FID* GC/MS
number t
R
(min) t
R
identification t
R
(min) MS + t
R
identification
8 40.162 n-undecane 10.391 linalool
cis-3-hexenylpropionate
δ-hexalactone
1-methyl-2,3-cyclohexadione
linalool
methyl benzoate
9 48.728 dihydrocarveol 14.750 n-decanal
methyl salicylate
estragole
n-decanal
octylacetate
10 71.366 anisylproprionate 27.134 valencene
valencene
piperonyl acetate
* For the GC/FID retention time identification, a 0.4-min window was used (±0.2 min)
Table 3 Continued
Another example of the GC/FID method appears
in Figure 2, showing the analysis of an Indonesian
nutmeg oil. Again a very detailed chromatogram is
obtained. Using the retention time locked database,
most constituents are identified. The small peak
eluting at 67.468 min is, for instance, identified as
isoeugenol. This is an important flavor compound.
010203040506070
67.468
80 90 min
Figure 2. GC/FID chromatogram of Indonesian nutmeg oil. (Conditions: Table 1)
6
0246 810121416182022242628min
10
9
8
7
6
5
4
1
3
2
GC/MS Method
For confirmation of solute identities and for the
identification of unknown peaks, the essential oils
are analyzed by GC/MS. The described method
uses a standard column and a faster temperature
program. These conditions are similar to the
method published by Adams [4]. The chromatogram
obtained for the Spanish orange oil is given in
Figure 3. A similar separation is obtained as in
Figure 1, but the resolution is lower due to the
lower column plate number. Moreover, some peak
overloading can be observed. Due to the fact that a
different temperature program is used, the com-
pounds also elute at different temperatures
(method not translated) and, therefore, also some
differences in relative elution profiles are observed
(see, for instance, relative elution of peaks 2, 3,
and 4). Nevertheless, this method can be used for
detailed identification of the essential oil con-
stituents. For this GC/MS method, a mass spectral
library was created containing approximately 400
compounds, together with the retention times.
With this library, identification is possible based
on mass spectra AND on retention time. The iden-
tification of the 10 main compounds (same as
labelled in Figure 1) using the Results Screener is
included in Table 3. The combination of mass spec-
tral and retention time data allows complete iden-
tification. It is also important to notice that the
correct compound was, in all cases, already
selected in the GC/FID retention time window.
Figure 3. GC/MS chromatogram of Spanish orange oil. (Conditions: Table 2, peak identification: Table 3)
7
The Indonesian nutmeg oil was also analyzed by
this GC/MS method. The chromatogram is given in
Figure 4. A classical library search using the stan-
dard NIST mass spectral library of the peak at
25.304 min gave isoeugenol as the first hit (proba-
bility 96%) and eugenol as the second hit (probabil-
ity 94%). The spectra of both compounds are very
similar (Figure 5). Using the Results Screener, the
compound was identified unequivocally as
isoeugenol (retention time plus mass spectral
match). This example clearly demonstrates the
power of combined retention time and mass
spectral search.
0 5 10 15 20 25 30 35 40 45 min
25.304
Figure 4. GC/MS chromatogram of Indonesian nutmeg oil. (Conditions: Table 2)
8
30 40 50 60 70 80 90 100 110 120 130 140 150 160 170
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
Abundance
72243 Eugenol
164
149
77
55 103 131
91 121
39
65
27
138
46
30 40 50 60 70 80 90 100 110 120 130 140 150 160 170
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
m/z
m/z
Abundance
273 Isoeugenol
164
149
77 103 131
91
55 121
65
39
32 138
110
Figure 5. Comparison of mass spectra of eugenol and isoeugenol.
Transformation of Retention Indices
Further it was evaluated if published retention
indices could be transferred into retention times
and if these calculated retention times match with
experimental data. A total of 34 test solutes were
analyzed. The compounds are listed in Table 4,
together with the FEMA code, the retention index
in an existing database (RI) [7], and the measured
retention time (t
R exp
) under retention time locked
conditions. From the retention index, the absolute
retention time was calculated using the retention
times of n-alkanes as reference compounds. The
calculated values are also listed in Table 4 (t
R calc
).
Thus, these retention times are not the original
retention times used for the retention index calcu-
lation, but calculated values. In the last column,
the difference between calculated and experimen-
tal retention times are also given. From these data,
it is clear that the calculated and experimental
retention times match very well (within ±0.2 min).
This means that retention times can be calculated
from the retention indices present in an existing
database using the locked retention times for
n-alkanes if the column dimensions and the tem-
perature program are the same. This is also valid
for the GC/MS method as shown in the following
example. Isoeugenol is present in the database of
9
FEMA t
R calc
t
R exp
t
r diff
Compound FEMA name code RI (min) (min) (min)
1 Acetal 2002 725.1 11.836 11.836 0.000
2 Amyl alcohol 2056 760.0 13.844 13.833 -0.011
3 Hexyl alcohol 2567 865.8 21.068 20.941 -0.127
4 Anisole 2097 923.0 25.529 25.403 -0.126
5 Ethyl acetoacetate 2415 944.4 27.300 27.319 0.019
6 Heptyl alcohol 2548 968.4 29.285 29.180 -0.105
7 Octanal 2797 1003.8 32.218 32.215 -0.003
8 Methyl 3-(methylthio)propionate 2720 1026.8 34.140 34.121 -0.019
9 Benzyl alcohol 2137 1036.9 34.984 35.011 0.027
10 Isoamyl butyrate 2060 1055.5 36.539 36.524 -0.015
11 Octyl alcohol 2800 1069.7 37.726 37.663 -0.063
12 Acetophenone 2009 1072.4 37.952 38.119 0.167
13 Benzylformate 2145 1081.2 38.688 38.851 0.163
14 Benzyl acetate 2135 1169.1 45.848 45,915 0.067
15 Allylheptanoate 2031 1180.0 46.729 46.664 -0.065
16 Decanal 2362 1207.4 48.917 48.970 0.053
17 Benzyl propionate 2150 1266.2 53.444 53.378 -0.066
18 1-Decanol 2365 1272.1 53.899 53.904 0.005
19 Anisyl alcohol 2099 1295.0 55.662 55.799 0.137
20 Isobornyl acetate 2160 1301.7 56.171 56.252 0.081
21 Benzyl isobutyrate 2141 1305.0 56.411 56.452 0.041
22 Undecanal 3092 1310.6 56.819 56.798 -0.021
23 Triacetin 2007 1344.4 59.279 59.192 -0.087
24 Benzyl butyrate 2140 1354.5 60.014 60.049 0.035
25 Acetanisole 2005 1369.0 61.070 60.960 -0.110
26 gamma-Nonalactone 2781 1373.4 61.390 61.273 -0.117
27 Anisyl acetate 2098 1426.0 65.116 65.203 0.087
28 Allyl cyclohexylpropionate 2026 1435.0 65.735 65.747 0.012
29 Lauryl alcohol 2617 1475.0 68.489 68.531 0.042
30 Isoamyl octanoate 2080 1487.0 69.315 69.230 -0.085
31 Isoamyl phenylacetate 2081 1503.0 70.405 70.387 -0.018
32 Ethyl-methylphenylglycidate 2444 1517.0 71.317 71.447 0.130
33 Ethyl-3-phenylglycidate 2454 1529.0 72.099 72.063 -0.036
34 gamma-Undecalactone 3091 1589.0 76.007 75.955 -0.052
Table 4. FEMA names, FEMA codes, retention indices, calculated retention times, experimental retention times, and retention
time differences for test solutes.
Adams [4] with a retention index of 1447. This
retention index can be transferred into an absolute
retention time using the following formula:
whereby: RI = retention index (from existing data
base),
Z = carbon number of preceding n-alkane,
t
RZ+1
and t
RZ
= retention times of following
and preceding n-alkanes (in RTL method)
and
t
RX
= retention time of solute in RTL
method
For isoeugenol, with an RI = 1447, and the preced-
ing tetradecane (Z=14) eluting at 23.26 min (t
RZ
),
and the following n-pentadecane (Z+1) eluting at
27.50 min (t
RZ+1
) using the retention time locked
GC/MS method, the calculated retention time is
25.25 min. This corresponds well (within ±0.2 min)
with the measured retention time (25.30 min).
Using these calculations, compounds can also be
added to the RTL databases.
Conclusion
Two methods were developed for the analysis of
essential oils. The first method is used for quality
control analysis. The method is locked using
n-pentadecane as locking standard. A retention
time locked database, containing approximately
[[RI – (Z ×100)] × (t
RZ + 1
_t
RZ
)]+ t
RZ
= t
RX
100
Agilent shall not be liable for errors contained herein or for incidental or consequen-
tial damages in connection with the furnishing, performance, or use of this
material.
Information, descriptions, and specifications in this publication are subject to change
without notice.
© Agilent Technologies, Inc. 2002
Printed in the USA
May 3, 2002
5988-6530EN
www.agilent.com/chem
For More Information
For more information on our products and services,
visit our Web site at www.agilent.com/chem.
400 compounds was created. This database can be
used to identify constituents based on their absolute
retention time under the locked conditions. The
locked method also guarantees retention time
stability in function of time, between columns and
between instruments.
Secondly, a GC/MS method was developed. This
method can be used for identification of essential
oil constituents. Identification is done based on the
combination of retention time and mass spectral
matching.
Finally, it is shown that retention indices for flavor
compounds measured under specific operational
conditions can be transferred into locked retention
times using the locked retention times of n-alkanes.
Thus, existing retention index databases can be
translated into locked retention time databases.
Moreover, a retention time locked database is not
restricted to the use of one (FID) detector, but
compounds detected by any GC detector can be
searched if the locked method is used.
References
1. R. Tabacchi and J. Garnero in P. Sandra and
C. Bicchi (Eds) , Capillary Gas Chromatography
in Essential Oil Analysis, Huthig, Heidelberg,
1987, pp 1-11.
2. W. Jennings and T. Shibamoto, Qualitative
Analysis of flavor and Fragrance Volatiles by
Glass Capillary Gas Chromatography, 1980,
Academic Press, New York.
3. T. Shibamoto in P. Sandra and C. Bicchi (Eds) ,
Capillary Gas Chromatography in Essential Oil
Analysis, Huthig, Heidelberg, 1987, pp 259-274.
4. R.P. Adams, Identification of Essential Oil
Components by Gas Chromatography - Mass
Spectroscopy, 1995, Allured Publishing
Corporation, IL, USA.
5. V. Giarrocco, B.D. Quimby and M.S. Klee,
Agilent Application Note 228-392, publication
(23) 5966-2469E, December 1997.
6. B.D. Quimby, L.M. Blumberg, M.S. Klee and
P.L. Wylie, Agilent Application Note 228-401,
publication (23) 5967-5820E, May 1998.
7. Quest International, Naarden, the Netherlands.