Talanta 51 (2000) 1179 1185

www.elsevier.com/locate/talanta

Continuous subcritical water extraction of medicinal plant

essential oil: comparison with conventional techniques
L. Gamiz-Gracia, M.D. Luque de Castro *
Analytical Chemistry Di6ision, Faculty of Sciences, Uni6ersity of Cordoba, E-14004 Cordoba, Spain
Received 13 October 1999; received in revised form 29 December 1999; accepted 29 December 1999

Abstract
A subcritical extractor equipped with a three-way inlet valve and an on/off outlet valve has been used for
performing subcritical water extractions (SWE) in a continuous manner for the isolation of the essential oil of fennel,
a medicinal plant. The target compounds were removed from the aqueous extract by a single extraction with 5 ml
hexane, determined by gas-chromatography-flame ionization (GC-FID) and identified by mass spectrometry (MS).
The proposed extraction method has been compared with both hydrodistillation and dichloromethane manual
extraction. Better results have been obtained with the proposed method in terms of rapidity, efficiency, cleanliness and
possibility of manipulating the composition of the extract. 2000 Elsevier Science B.V. All rights reserved.
Keywords: Subcritical water; Gas chromatography; Essential oil; Medicinal plant

1. Introduction
Fennel (Foeniculum 6ulgare) has been used traditionally in medicine for the treatment of several
stomach affections (as aerophagia or diarrhea)
and obesity. There are also commercial pharmaceuticals with formula based on fennel essential
oil. The essential oil from plants has usually been
isolated by either steam distillation [1 3] or solvent extraction [4 6]. These techniques present
some shortcomings, namely losses of volatile compounds, low extraction efficiency, long extraction
* Corresponding author. Tel.: +34-957-218614; fax: + 34957-218606.
E-mail address: qa1lucam@uco.es (M.D. Luque de Castro)

time, degradation of unsaturated compounds and

toxic solvent residue, that make mandatory the
use of alternative techniques for the extraction of
essential oils [7].
Continuous subcritical water extraction is a
technique based on the use of water as extractant,
at temperatures between 100 and 374C and pressure high enough to maintain the liquid state.
This technique is emerging as a powerful alternative for solid sample extraction [8]. Its use in the
field of essences is recent and very promising
[9,10] and it appears as a useful alternative to
conventional and supercritical CO2 extractions
[11].
The aim of this work was to develop a method

0039-9140/00/$ - see front matter 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 0 3 9 - 9 1 4 0 ( 0 0 ) 0 0 2 9 4 - 0

1180

L. Gamiz-Gracia, M.D. Luque de Castro / Talanta 51 (2000) 11791185

for the continuous subcritical water extraction of

medicinal essential oils, and compare the results
with those obtained by conventional techniques,
in order to introduce this advantageous alternative in the pharmaceutical field.

2. Experimental

2.1. Instruments and apparatus

The subcritical water extractions were performed using the following assembly: a Shimadzu
LC10AD high pressure pump was used to propel
the extractant water through the system. The extractor, described elsewhere [7] (a prototype designed and patented by Salvador and Merchan
[12]) consisted of a stainless steel cylindrical extraction chamber (15011 mm id, 14 ml internal
volume) and it was closed with screws at both
ends in order to permit the circulation of the
leaching fluid through it. The screw caps also
contained stainless steel filter plates (2 mm in
thickness and 1/4 in id) to ensure that the plant
material remained in the extraction chamber. This
chamber, together with a stainless steel preheater,
was located in an oven, designed to work at up to
300C and its temperature was controlled using a
Toho TC-22 temperature controller. A loop made
from a 1 m length stainless steel tubing
and cooled with water at room temperature, was
used to cool the fluid carrying out from the oven
to a temperature close to 25C, thus avoiding
losses of volatile components caused by the hot
water.
A simple laboratory quickfit apparatus
which contained a 2000 ml steam generator
flask, a distilling flask, a condenser and a receiving vessel was used to perform the hydrodistillation.
The analyses of the extracts were all performed
by using a Varian Star 3400 gas chromatograph
equipped with a SGL-5 fused silica column (25
m 0.25 mm id, 0.25 mm film thickness) and a
flame ionization detector (FID). Finally, a Fison
VG Autospec (Micromass Instruments) mass
spectrometer was used to identify the compounds
in the extracts.

2.2. Reagents
Fennel (F. 6ulgare) was collected in the South
of Spain (Cordoba). A stock standard solution of
5200 mg ml 1 of nonane (Sigma, St Louis, MO)
in HPLC grade hexane (Scharlau, Barcelona,
Spain) was prepared. For the liquidliquid extraction step of the aqueous extracts NaCl,
Na2SO4 (Merck, Darmstadt, Germany) and
HPLC grade hexane were used as demulsifier,
drying agent and extractant, respectively. Bidistilled degassed water purified through a Milli-Q
deionizing unit (Millipore) and dichloromethane
(Prolabo, France) were used as extractants.
All the gases were of 95% purity or higher
(Carburos Metalicos, Barcelona, Spain).

2.3. Sample preparation

Fennel was stored at 4C until analysis. A 4.0-g
sample was cut immediately prior to subcritical
water, hydrodistillation or organic solvent extraction, in order to avoid losses of volatile
components.

2.4. Procedure
2.4.1. Subcritical water extraction (SWE)
The cell was filled with fennel cut in small
pieces and weighed (4.0-g), and glass wool plugs
were inserted in both ends of the cell to prevent
the frit from being plugged. After assembling the
extraction cell in the oven, this was brought up to
the working temperature (150C) and pressurised
with ca. 50 bar of water by opening the inlet
needle valve from the pump. The valve was then
closed and static extraction was developed for 30
min. The inlet and outlet valves were then opened,
water was pumped through the chamber at a flow
rate of 2 ml min 1 and the extract was collected
in a vial at room temperature. For kinetic experiments, the vial was replaced at pre-set intervals of
5 min (10 ml each).
2.4.2. Hydrodistillation
A 4.0-g portion of fennel was subjected to
hydrodistillation in the assembly described above.
The steam generator flask was filled with 350 ml

L. Gamiz-Gracia, M.D. Luque de Castro / Talanta 51 (2000) 11791185

of Milli-Q water and heated with a heating

mantle. As the water vaporized, the steam passed
to the distillation flask containing the fennel. The
vapour then passed through the cooled tube
where it condensed. The distillate (94 ml) was
finally collected into the receiving flask after 4 h
distillation.

2.4.3. Liquid liquid extraction

After subcritical and hydrodistillation extractions, a liquid liquid extraction step was performed, in which volumes of 5 ml of hexane and
65 ml of nonane stock solution (used as extractant
and internal standard, respectively) were added to
each extract in a separatory funnel, and ca. 0.5 g
of NaCl was added to facilitate the breaking of
the emulsion. The organic extract was then dried
with anhydrous Na2SO4.
2.4.4. Dichloromethane extraction
A 4.0-g portion of fennel was placed in a glass
screw-top vial and 130 ml of internal standard and
10 ml of dichloromethane were added to the
sample. The extraction was performed on a
shaker for 24 h [13].
2.4.5. Chromatographic separation and detection
Aliquots of 2 ml of hexane after either SWE or
hydrodistillation plus liquid liquid extraction in
Table 1
Optimisation of variables
Variable
SWE
Temperature (C)
Pressure (bar)
Flow rate
Amount of sample (g)

Range studied

50200

0.53.0
2.05.0

Gas chromatography
Carrier
Injection volume (ml)
Column
Split ratio
1:30/1:6
Carrier flow rate (ml
min1)
Temperature gradient (C min1)
Ramp 1
15
Ramp 2
15

Optimum value

150
20
2.0
4.0
Helium
2
SGL-5
1:6
1.0

1
5

1181

both cases, and of 2 ml of dichloromethane from

the manual extraction, were injected into an SGL5 fused silica column (25 m0.25 mm id, 0.25
mm film thickness). The carrier gas (helium) was
delivered at a flow rate of 1.0 ml min 1. The
injector and detector temperatures were 250 and
300C, respectively. The oven temperature was
55C for 2 min, then increased to 75C at a rate
of 1C min 1, kept for 1 min, and finally increased to 250C at a rate of 5C min 1. The split
ratio was 1:6.

2.4.6. Gas chromatography-mass spectrometry

identification
Fractions of the hexane extracts obtained by
subcritical fluid extraction and hydrodistillation
were collected in order to identify the components
by gas chromatography-mass spectrometry. Mass
spectra (electron impact) of the compounds were
obtained using a Fison VG Autospec (70 eV) by
direct injection of 2 ml of the extract into the
ionization chamber at a temperature of 250C.

3. Results and discussion

A static-dynamic approach was tested for the
isolation of fennel oil and then compared with
conventional techniques such as hydrodistillation
and organicsolvent extraction. A static extraction period (for enhancing sampleextractant
contact thus favoring the attainment of the partition equilibrium) was carried out, followed by a
dynamic extraction period in which fresh extractant passed continuously through the extraction
chamber, thus displacing the equilibrium to
quantitativeness.

3.1. Optimisation of 6ariables

The variables affecting the subcritical extraction
were studied in order to maximise the yield of
essential oil in as short a time as possible. With
this aim the univariate method was used. The
static period was kept constant at a value of 5
min. The optimum values found for the variables
are shown in Table 1.

1182

L. Gamiz-Gracia, M.D. Luque de Castro / Talanta 51 (2000) 11791185

Fig. 1. Effect of the temperature in the main components of essential oil from fennel.

Fig. 2. Effect of the flow rate in the main components of essential oil from fennel. IS= internal standard.

The temperature of the extraction chamber was

the key variable affecting continuous subcritical
extraction. The yield increased with temperature up
to 150C. At 175C, the yield of most of the
monoterpene compounds increased, but decreased
for the oxygenated compounds, probably due to
their degradation (Fig. 1). For temperatures higher
than 175C, a strong burning smell of the extract
appeared. A value of 150C was selected as optimum.
The pressure, as demonstrated in previous works
[10], had little influence on the yield of essential oil,
as long as the extractant water was under liquid
state. Thus, a minimum pressure of 20 bar was
maintained throughout all extractions.
The flow rate was studied in the interval 0.53.0
ml min 1. Increased flow rates enhanced the extraction of most of the monoterpene compounds,
while in the case of oxygenated compounds the
extraction rate was maximum at 2 ml min 1, as can

be seen in Fig. 2. A flow rate of 2 ml min 1 was

selected for further experiments.
The influence of cutting or grounding the sample
was also studied. When the sample was ground, the
yield of the essential oil decreased, probably due to
losses of volatiles during preparation. Thus, for
further experiments, the sample was just cut into
small pieces (ca. 1 cm in length).
The amount of sample was also studied in the
interval 25 g (the maximum capacity of the cell),
and 4 g was selected as it provided signals high
enough.
In the manual liquidliquid extraction, the number of extraction steps was studied by performing
a second 5 ml hexane extraction to ensure that no
compounds remained in the aqueous phase after
the first extraction. Only small peaks of the majority compounds were obtained after injection of the
second extract, so a single 5 ml hexane extraction
was carried out in subsequent experiments.

L. Gamiz-Gracia, M.D. Luque de Castro / Talanta 51 (2000) 11791185

3.2. Effect of the static extraction time on the

o6erall extraction kinetics
The influence of the static extraction time on
both the extraction kinetics and composition of
the oil was studied by performing extractions
with subcritical water consisting of a static
extraction step of 0, 5, 15 and 30 min followed
by 20 min of dynamic extraction (at 2 ml

1183

min 1) during which the vial was replaced

every 5 min (ca. 10 ml). As can be seen in
Fig. 3, increased static extraction times
enhance the extraction of most of the compounds.
It can also be established that for the monoterpene compounds, 20 ml is enough to complete the
extraction, while in the case of the oxygenated
ones, at least 40 ml of subcritical water is required.

Fig. 3. Effect of the static extraction time in the main components of essential oil from fennel.

Fig. 4. Chromatograms of the SWE (a), dichloromethane extraction (b) and hydrodistillation (c) of fennel essential oil. Peaks
identification: (1) a-pinene; (2) b-pinene; (3) b-myrcene; (4) a-phelandrene; (5) limonene; (6) unknown; (7) camphor; (8) terpinen-4ol; (9) linalyl propanoate; (10) unknown; (11) anethol; (12): copaene.

1184

L. Gamiz-Gracia, M.D. Luque de Castro / Talanta 51 (2000) 11791185

Table 2
Comparison of area compound/internal standard ratio for GC of extracts obtained after hydrodistillation, dichloromethane
extraction and SWE
Compound

Hydrodistillationa

Dichloromethane extractionb

SWEa

a-Pinene
b-Pinene
b-Myrcene
a-phelandrene
Limonene
Unknown
Camphor
Terpinen-4-ol
Linalyl propanoate
Unknown
Anethol
Copaene

0.265
0.067
0.569
2.397
5.858
25.196
0.697

0.154
5.857
25.214

0.499
0.140
0.667
2.580
5.576
3.529

1.272
5.887

4.889
0.887
5.469
18.689
51.598
67.732
1.503
0.225
0.781
23.613
68.054
0.234

a
b

Extracted in 5 ml of hexane.
In 10 ml of dichloromethane.

3.3. Manipulation of the oil composition

As a consequence of the relationship between
the extraction conditions and qualitative composition of the extract, the latter can be manipulated
by changing the conditions of the extraction (temperature, flow rate and static extraction time).

3.4. Precision of the SWE/chromatographic

method
The precision (expressed as RSD) of both the
chromatographic step and the overall process (extraction and detection) was studied separately.
The precision of the chromatographic step was
calculated by injecting one of the organic extracts
from SWE of 4.0 g of fennel seven times. The
average RSD for the eight major components was
3.26%. For studying the precision of the overall
process, 5 portions of 4.0 g of fennel were subjected to SWE under the same conditions and the
hexane layer obtained after liquid liquid extraction was injected in the GC. The resulting RSD
for the eight major components was 6.77%.

3.5. Comparison of methods

The SWE method (30 min of static extraction

followed by 20 min of dynamic extraction at 2 ml

min 1) has been compared with hydrodistillation
and manual dichloromethane extraction as conventional alternatives in terms of rapidity, efficiency, cleanliness and possibility of manipulating
the composition of the extract.

3.5.1. Rapidity
The SWE is clearly quicker than its conventional counterparts. The extraction takes 50 min,
whilst 4 and 24 h were required by hydrodistillation and manual dichloromethane extraction,
respectively.
3.5.2. Efficiency
The most efficient method in terms of quantitative composition of the extract seems to be SWE,
as can be seen in the chromatograms of the
extracts obtained by the three methods for the
same amount of sample, and the values of the
area ratios for the main component and the internal standard (Fig. 4 and Table 2, respectively). It
must be taken into account that in the case of
SWE and hydrodistillation the final extracts was
obtained in 5 ml of hexane, and in the case of
manual dichloromethane extraction, the final volume was 10 ml.

L. Gamiz-Gracia, M.D. Luque de Castro / Talanta 51 (2000) 11791185

3.5.3. Cleanliness
SWE is a very clean method, which avoids
both the use of large amounts of organic solvents and residue generation.
3.5.4. Possibility of manipulating the composition
of the extract
The SWE allows the manipulation of the
composition of the extract by changing extraction parameters such as temperature, flow rate
and/or static extraction time. This is unfeasible
with the conventional methods.

4. Conclusion
The continuous SWE of essential oil in medicinal plants (fennel) has been studied. The
method offers important advantages over traditional alternatives, namely: shorter extraction
times (50 min against 4 h for hydrodistillation
and 24 for manual dichloromethane extraction);
cost (energy cost is fairly higher for performing hydrodistillation than that required for
reaching subcritical conditions); cleaner features
(as no a great volume of organic solvent is involved); and the possibility of manipulating the
composition of the oil by changing the parameters of the extraction (temperature, flow rate
and static extraction time). The method contributes to the automation of pharmaceutical industry.

1185

Acknowledgements
The Spanish Comision Interministerial de
Ciencia y Tecnologa (CICyT) is thanked for
financial support (Project no. PB96-1265). The
Department of Organic Chemistry (Faculty of
Sciences, University of Cordoba) is thanked for
realising the GC-MS identifications.
References
[1] J.C. Chalchat, R.P. Garry, A. Michet, Flavour Fragr. J. 6
(1991) 189.
[2] M.D. Guillen, N. Cabo, J. Murillo, J. Sci. Food Agric. 70
(1996) 359.
[3] H.L. De-Pooter, E.A. Aboutabl, A.O. El-Shabrawy,
Flavour Fragr. J. 10 (1995) 63.
[4] R.K. Verma, G.C. Uniyal, M.M. Gupta, Indian J.
Pharm. Sci. 52 (1990) 276.
[5] M. Taverna, A.E. Baillet, D. Baylock, J. Ass. Off. Anal.
Chem. 73 (1990) 206.
[6] S.K. Volkov, E.I. Grodnitskaya, J. Chromatogr. 769
(1997) 275.
[7] M.D. Luque de Castro, M.M. Jimenez Carmona, V.
Fernandez Perez, Trends Anal. Chem. 18 (1999) 708.
[8] M.D. Luque de Castro, M.M. Jimenez Carmona, Trends
Anal. Chem. 17 (1998) 441.
[9] A. Basile, M.M. Jimenez Carmona, A. Clifford, J. Agric.
Food Chem. 46 (1998) 5205.
[10] M.M. Jimenez Carmona, J.L. Ubera, M.D. Luque de
Castro, J. Chromatogr. 855 (1999) 625.
[11] S.H.R. Al-Saidi, A. Clifford, A. Basile, Fresenius J. Anal.
Chem. (1999) submitted for publication.
[12] USA Patent no. 5, 400, 642, 1995.
[13] J.M. Bowman, M.S. Braxton, M.A. Churchill, J.D. Hellie, S.J. Starrett, G.J. Ellis, S.D. Ensley, S.J. Maness, C.D.
Meyer, J.R. Seller, Y. Hua, R.S. Woosley, D.J. Butcher,
Microchem. J. 56 (1997) 10.