Available online at www.sciencedirect.

com

J. of Supercritical Fluids 45 (2008) 3742

Supercritical fluid extraction of bioactive compounds from

sunflower leaves with carbon dioxide and water on a pilot plant scale
L. Casas a, , C. Mantell a , M. Rodrguez a , A. Torres
b
, F.A. Macas b , E.J. Martnez de la Ossa a
a

Department of Chemical Engineering, Food Technology and Environmental Technologies, Faculty of

Science, University of Cadiz, Box 40, 11510 Puerto Real, Cadiz, Spain
b
Department of Organic Chemistry, Faculty of Science, University of Cadiz,
Box 40, 11510 Puerto Real, Cadiz, Spain
Received 14 June 2007; received in revised form 16 November 2007; accepted 7 December 2007

Abstract
An analysis of the supercritical fluid extraction process on Helianthus annuus L (sunflower) using carbon dioxide and water on a pilot plant
scale has been carried out. The possibility of using the resulting extract as an herbicide was analyzed by bioassays. The influence of time for
different solvent flow rates was investigated. The extraction pressure and temperature were optimized in previous studies and a pressure of 380
bar and a temperature of 50 C were selected. The bioactivities and extraction yields obtained under the different conditions were compared.
The best activity profiles were obtained for 40 g/min CO2 + 1 g/min H2 O and 80 g/min CO2 + 2 g/min H2 O.
2008 Elsevier B.V. All rights reserved.
Keywords: Sunflower; Supercritical carbon dioxide; Extraction; Bioactive compounds

1. Introduction
The traditional methods for the extraction of plant materials include steam distillation and organic solvent extraction
using percolation, maceration or Soxhlet techniques. These
procedures, however, have distinct drawbacks such as timeconsuming and labour-intensive operations, handling of large
volumes of hazardous solvents and extended concentration
steps that can result in the loss or degradation of target
analytes. Moreover, there is increasing interest in alternative
extraction technologies that consume smaller quantities of
organic solvent because of the rising solvent acquisition and
disposal costs and regulatory restrictions [1].
Supercritical fluids have been shown to exhibit several
advan- tages in the extraction of natural products from plant
matrices. The combined liquid-like solvating capabilities and
gas-like transport properties of supercritical fluids make them
particu- larity suitable for the extraction of diffusion-controlled
matrices such as plant tissues. Moreover, the solvent strength
of a super-

Corresponding author. Tel.: +34 956 016579; fax: +34 956 016411.
E-mail address: lourdes.casas@uca.es (L. Casas).

0896-8446/$ see front matter 2008 Elsevier B.V. All rights

reserved. doi:10.1016/j.supflu.2007.12.002

critical fluid can be easily tuned by simply changing the

applied pressure and/or temperature.
Carbon dioxide, the most commonly used supercritical
fluid, has the advantages of being non-flammable, fairly nontoxic, cost-effective and easily removed from the extract
following decompression. Finally, due to its relatively low
critical temperature (31.1 C), thermal sample decomposition is reduced.
Pure
CO2 , however, is not an appropriate extraction fluid for polar
analytes and retentive matrices. In order to enhance the solvating power of CO2 , the addition of a small amount of a
modifier solvent is required [2].
Supercritical fluid extraction (SFE) can be applied to
systems on a range of different scales, for instance from an
analytical scale (less than a gram to a few grams of sample) to
pilot plant scale (several hundred grams or kilograms of
sample) and even up to large industrial scale (tons of raw
material).
In previous studies we began to address the extraction of
bioactive substances from sunflower leaves at the analytical
scale (10 ml extraction vessel). Studies were carried out on
the pre- treatment of samples [3] and the effect of the
presence of a cosolvent (water, methanol and dimethyl
sulphoxide (DSMO)) on both the yield and the bioactivity of
the resulting extracts [4]. The use of methanol, water and
DMSO as modifiers increased

L. Casas
L. Casas
et al.
et al.
/ J./ofJ. Supercritical
of Supercritical
Fluids
Fluids
45 45
(2008)
(2008)
3742
3742

the extraction efficiency of the process but the best yields were
obtained using water as a modifier. Furthermore, the results
obtained show that the properties under investigation
(pressure, temperature and pre-treatment of the samples) do not
have a sig- nificant influence on the bioactivity profiles of the
products. The dry sample extracted at 500 bar with CO2 + 5%
H2 O gave the best yields and a satisfactory activity profile. It
is therefore of interest to continue to develop the process using
carbon dioxide with water as a cosolvent.
The work described here involved the extraction of bioactive compounds from sunflower leaves (Helianthus annuus L)
(both congealed samples and dried samples) with supercritical
carbon dioxide and water on a pilot plant scale (2 l extraction
vessel). The aim of the study was to assess the effect that the
flow rate, percentage of water and extraction time had on the
supercritical extraction process. The extraction pressure and
temperature were optimized in previous studies and values of
380 bar (pressure) and 50 C (temperature) [3] were selected,
with water chosen as the cosolvent [4]. The extraction yield
and the bioactivity of the extracts were analyzed.
2. Materials and methods
2.1. Samples and chemicals
Leaves of H. annuus L (variety Aitana) were collected in
July
2005 during the third plant development stage [5] (plants were
1.2 m tall with flowers, 1 month before harvest) and plants
were provided by Rancho de Merced, Agricultural Research
Station (CIFA), Junta of Andaluca, Jerez, Spain.
The sample was stored under two sets of conditions in order
to evaluate the behaviour of each in terms of extraction yield
and bioactivity of the extracts:
sample congealed at 25

C,
sample dried at room temperature (25 C 1 C) until a
constant weight was reached.
The specifications of the other chemical reagents used are
given in Table 1.
2.2.
Extraction
pressure

at

high

Fig. 1. Schematic diagram of the equipment used for the SFE.

maximum flow rate of 150 g/min of carbon dioxide and 50

g/min of cosolvent. A thermostated jacket allowed control of
the extraction temperature. The cyclonic separator allowed
period- ical discharge of the extracted material during the SFE
process. A schematic diagram of the SFE apparatus used in
this research is shown in Fig. 1.
The operating methodology involved loading the extraction cartridge with approximately 190 g of the sample, which
had previously been homogenized in order to maintain a constant apparent density in all experiments. The valves MV-1
and MV-2 were closed and the valves MX-1 and MV-4 were
opened.
When the heat exchanger (HE1) was at a temperature of
approximately 1 C the CO2 cylinder was opened. All of the
operating conditions apart from the flow of CO2 were
established and when the system had stabilized the extractor
was pressurized with CO2 . When the pressure in the extractor was near to
the desired extraction pressure valve MV-2 was opened to start

The extractions were carried out in a pilot plant from Thar

Technology (Pittsburgh, PA, USA, model SF2000) provided
with an extraction vessel (capacity of 2 l) and two pumps with
a

Table 1
Reagents
Reagent

Purity

Source

Use

Carbon dioxide
Water
Acetone
Citric acid monohydrate

99.995%
Milli Q
PA
PA

Carburos metalics

Extraction at high pressure

Cosolvent
Collection of extracts
Preparation of buffer

Panreac
Panreac

Potassium phosphate di-basic 3-hydrate

Sucrose
Dimethyl sulfoxide

L. Casas
L. Casas
et al.
et al.
/ J./ofJ. Supercritical
of Supercritical
Fluids
Fluids
45 45
(2008)
(2008)
3742
3742
PA
Panreac
PA
Panreac
PA
Panreac

Preparation of buffer
Preparation of buffer
Dissolution of the extracts

the flow of solvent. When a balanced state had been attained,

BPR1 was opened automatically and the process began.
The extracts were collected in S1 and acetone was added
to dissolve the extract. Valve MV-3 was subsequently used to
collect the extracts in glass bottles and these were stored at 4

C
with the exclusion of light. The acetone was later evaporated
and the resulting extract weighed.
The experiments on each sample were carried out in
duplicate in order to evaluate the variability of the
measurements.
The extraction pressure and temperature were optimized in
previous studies and values of 380 bar (pressure) and 50 C
(tem- perature) were used. The extraction times selected
were: 30,
60, 90, 120, 180 and 300 min at 25 g/min CO2 + 1 g/min H2
O,
40 g/min CO2 + 1 g/min H2 O, 40 g/min CO2 + 2 g/min H2
O,
80 g/min CO2 + 2 g/min H2 O and 80 g/min CO2 + 4 g/min H2
O.
2.3. Coleoptile bioassay
Bioassays constitute an important tool to evaluate the
inhibit- ing or stimulating activity in terms of growth of the
substances extracted according to the conditions described in
the previous section.
Wheat seeds (Triticum aestivum L cv. Duro) were sown in
15 cm diameter Petri dishes moistened with water and were
grown in the dark at 22 1 C for 3 days. The roots and caryopses were removed from the shoots. The latter were placed in
a guillotine and the apical 2 mm sections were cut off and discarded. The next 4 mm lengths of the coleoptiles were
removed and used for bioassays. All manipulations were
performed under a green safelight. Compounds were dissolved
in DMSO and diluted to the final bioassay concentration.
Parallel controls were also run [5].
A sample (16 mg) of each extract obtained under the conditions described in Section 2.2 was weighed out. The extracts to
be assayed for biological activity were added to test tubes and
were dissolved in 16 ml of an aqueous solution of
phosphate/citrate buffer (pH 5.6) containing 2% sucrose. The
extracts were insolu- ble in water and DMSO (5 l per ml of
plug) was therefore added to ensure total dissolution. Solutions
of 500, 250 and 125 ppm were prepared in a similar way for
each extract.
Five coleoptiles were placed in each test tube and the
samples were rotated at 6 rpm in a roller tube apparatus for 24
h at 22 C in the dark. The coleoptile lengths were measured
by digitalization
of their images. Data are presented as percentage differences
from the control.
Each assay was performed four times and on two different
days.
2.4. Cluster analysis
Statistical treatments were performed using the SPSS 10.0
program (Statistical Package for Social Sciences). Association

analysis of the data based on the bioactivity profile was performed for each of the different sets of extraction conditions.
To further clarify the relationships between the clusters and
those individuals forming the clusters, a dendrogram was generated by hierarchical cluster analysis; the squared Euclidean

distance between normalised data was used to measure the

sim- ilarity between samples.
3. Results and discussion
3.1. Extraction yields
Interesting extraction characteristics are often obtained on
running the extraction experiment at different flow rates. In
order to determine the controlling stage of the extraction
process, it is necessary to study the effect of the solvent flow
rate. If the solvent flow rate is relatively low, the controlling
stage will be the solubility of the solute in the solvent. If the
solvent flow rate is relatively high, then the controlling stage
will be the diffusion of the solvent into the matrix.
The extraction yields expressed as mg of extract/100 g of
dry leaves are shown in Fig. 2 for different extraction times
and under different conditions in terms of pre-treatment of
sample, solvent flow rates and solventCO2 mixtures. Each
experimental point on the extraction curve represents the
averages of the two inde- pendent experiments with a
reproducibility of approximately 5% CV (coefficient of
variation).
It can be seen that the recovery is very dependent on the
pre-treatment. According to our experimental data (Fig. 2), the
best extraction yields were obtained from the dried samples.
The moisture from the congealed samples seems to be a factor
that diminishes the extraction yield, with the water acting as a
Fig. 2. Extraction yields for different extraction times.

solvent that competes with supercritical CO2 . If excess water

remains in the extraction vessel, highly water-soluble solutes
prefer to partition into the aqueous phase and, consequently,
the SFE recovery will be low.
It can also be seen that the recovery is very dependent on
the flow rates. The experimental results indicate a higher
extrac- tion yield for the process with a solvent flow rate of
80 g/min CO2 + 4 g/min H2 O than that obtained at 25 g/min
CO2 + 1 g/min H2 O. Generally, an increase in the flow rate
leads to higher yields of extracts. This effect is due to the
presence of larger amount of solvent in the operation, a factor
that improves the extraction yield.
The addition of modifiers to supercritical CO2 not only
enhanced the solubility of the analyte in the fluid but also
decreased the interaction of the analyte and the matrix. In this
way, different extraction results can be obtained by
manipulating the ratios of liquid modifier.
The experimental results obtained indicate that the most
appropriate extraction time is approximately 120 min for all of
the selected flow rates. Above this range the increase in the
extraction yield is not significant and therefore there is little
advantage in prolonging the extraction.
The increase in the extraction yields for different flow rates
was more marked for the dried samples than for the congealed
samples and the best extraction yields were achieved using
dried samples. For the congealed samples an increase in the
flow rates to 80 g/min CO2 and 2 or 4 g/min H2 O did not
lead to an improvement in the extraction process, despite the
fact that solvent consumption was doubled. On the basis of
these results, it is not advisable to carry out the extraction at
a flow rate of
80 g/min CO2 . However, it is worth noting that the increase
in extraction yield obtained with dried samples on increasing
the flow rate is much more marked than for congealed
samples.
In both cases, on using flow rates of 40 and 80 g/min an
increase in the percentage of water present in the solvent
system did not give rise to significant increases in the yields.
For this reason it was appropriate to carry out general
biological assays to select the best flow rates and extraction
times in the process. Other aspects, such as economic
considerations, must also be taken into account when choosing
appropriate parameters for a process of this type.
3.2. Bioassay
It was necessary to perform a general bioassay in order to
select the conditions (extraction times, solvent flow rates and
solventCO2 mixtures) that provide the extracts with the best
bioactivities. The aim of this study was not to determine
specific values, but to attempt to obtain activity profiles on the
basis that an extract will be more bioactive when its activity
levels persist as the sample is diluted.
The results of the bioactivity assays are shown in Figs. 3
and 4 for congealed and dried samples, respectively. The data
are expressed as percentage differences from the control,
which means that a value of zero represents an identical value
to the control. On the other hand, a positive value represents
stimula- tion of the parameter in question and a negative value
represents

Fig. 3. Bioactivities of extracts obtained from congealed samples (extraction

time/CO2 flow + H2 O flow).

inhibition of the growth of the wheat coleoptiles under the

given experimental conditions. The experimental error was
approxi- mately 8.4%.
It can be seen from Fig. 3 that the extracts obtained from
con- gealed samples at 25 g/min CO2 + 1 g/min H2 O do not
show any difference in terms of the activity levels of the
extracts obtained with different extraction times; at 1000 ppm
the extracts showed an activity of around 63% and values
near to 20% for the
75-ppm dilution. According to the cluster analysis (Fig. 5)
these assays are united in a sub-group.
It can be seen from Fig. 4 that the activity profiles of the
extracts obtained at 25 g/min CO2 + 1 g/min H2 O from
dried samples are also very similar for different extraction
times; at
1000 ppm the extracts showed 63% activity and values near
to
10% for the 75 ppm dilution. According to the cluster
analysis
(Fig. 5) these assays also form a well-defined sub-group.
On the basis of the results described above, it can be
concluded that variations in extraction times do not cause
appre- ciable changes in the bioactivities of the samples at low
flow rate. Furthermore, the congealed samples give rise to
better activity profiles than the dried samples. The most
appropriate extraction

Fig. 4. Bioactivities of extracts obtained from dried samples (extraction

time/CO2 flow + H2 O flow).

found in this study, i.e. only a small decrease in the activity on

increasing the dilution.

Fig. 5. Cluster analysis (where S is for dried sample and C is for congealed
sample).

time can be defined according to the yields and not the activity
profiles.
In the case of the congealed samples extracted for 120 min,
an increase in flow to 40 g/min CO2 with a constant flow rate
of water (1 g/min) improves the activity of the extract because
the activity level does not decrease drastically with dilution.
How- ever, an increase in extraction times at 40 g/min CO2 +
1 g/min H2 O proved detrimental to the activity of the extract.
The activ- ity profiles of the extracts obtained at 40 g/min
CO2 + 2 g/min H2 O from congealed samples are also proved
detrimental to the activity of the extract with an increase in
extract time. Compar- ison of the activity profiles of the
extracts obtained with these two flow rates shows that an
increase in the percentage of water caused a decrease in the
activity profile (Fig. 3).
It can be seen from Fig. 5 that for dried samples an increase
in the percentage of water also leads to a decrease in the
activity profile of the extracts. At this point we must consider
that studies previously [4] the flow of solvent was down and
in the current study did above. Because working conditions of
water are denser than that of carbon dioxide and that part of
the bioactive sub- stances are soluble in water, they are left
accumulated in the exhaust, thus decreasing the bioactivity
profiles of the extracts to increase percentage of water. At 40
g/min CO2 + 1 g/min H2 O the extracts show a small decrease
in activity on increasing the dilution; 1000 ppm shows 100%
activity and 125 ppm shows
45% activity. Nevertheless, at 40 g/min CO2 + 2 g/min H2
O,
1000 ppm shows inhibition levels of 78% and this decreased
significantly to 38% at a dilution of 500 ppm.
It can be seen from the cluster analysis in Fig. 5 that
there is a clear difference between the activity levels of the
extracts obtained with different pre-treatment regimes and
flows of 40 g/min CO2 + 1 g/min H2 O and CO2 + 2 g/min H2
O. The extracts obtained at 40 g/min CO2 + 1 g/min H2 O
from dry samples and those obtained from congealed samples
with an extraction time of 120 min and the same flow show
the best bioactivity profiles. These assays form a
conglomerate that cor- responds to the best activity profiles

In the extraction of congealed samples conditions of 80

g/min CO2 + 4 g/min H2 O are extremely unfavourable to
the pro- cess, as evidenced by the marked decrease in the
activity level on dilution. An increase in the extraction
times at 80 g/min CO2 + 2 g/min H2 O and 80 g/min CO2 + 4
g/min H2 O gave rise to a marked decrease in the activity level
on dilution.
The cluster analysis shown in Fig. 5 indicates that dried
sim- ples extracted at 80 g/min CO2 + 2 g/min H2 O also
formed a conglomerate with the best activity profiles, i.e.
only a small decrease in the activity on increasing the dilution.
When the process was performed with congealed samples,
both the extraction time and the percentage of water were
found to have a significant influence on the bioactivity of the
extracts. Nevertheless, when the process was performed with
dried sam- ples the number of significant variables was
reduced and only the percentage of water is a significant
factor for the bioactivity of extracts. Therefore, low water
contents and extraction times of 120 min are the best
conditions to obtain the extracts because
despite the fact that they do not give rise to the maximum
yield
these extracts do show the best bioactivity profiles.

4. Conclusions

References

The study described above allows the following

conclusions to be drawn concerning the yield of extract:

[1] Q. Lang, C.M. Wai, Supercritical fluid extraction in herbal and natural
product studiesa practical review, Talanta 53 (2001) 771782.
[2] J. Liu, B. Han, G. Li, Z. Liu, J. He, G. Yang, Solubility of the non-ionic
surfactant tetraethylene glycol n-laurel ether in supercritical CO2 with npentanol, Fluid Phase Equilib. 15 (2001) 247254.

The extraction yields from congealed samples are low

com- pared to those obtained from previously dried samples
under a given set of conditions of flow rate, extraction time
and percentage of water.
It is not advisable to work with a flow rate above 25 g/min
CO2 in the case of congealed samples because the increased
solvent consumption is not compensated by increases in the
yields.
From dried samples it is advisable to work at 80 g/min CO2
on the basis that the extraction yields are higher. Nevertheless, significant differences in the extraction yields were not
detected for different flow rates of water.
As far as the biological activity is concerned, the extraction
times and percentage of water present in the solvent system
have an influence on the bioactivity of the extracts:

The best activity profiles are obtained with an extraction

time of 120 min.
An increase in the percentage of cosolvent, at constant
flow rates (40 g/min CO2 and 80 g/min CO2 ), proved
detrimental to the extraction process in terms of the
bioactivity of the extracts.

[3] L. Casas, C. Mantell, M. Rodrguez, M.D. Gordillo, A. Torres, F.A. Mac

as, E. Martnez de la Ossa, Effect of the pre-treatment of the samples
on the natural substances extraction from Helianthus annuus L. using
supercritical carbon dioxide, Talanta 67 (2005) 175181.
[4] L. Casas, C. Mantell, M. Rodrguez, A. Torres, F.A. Macas, E. Martnez
de la
Ossa, Effect of the addition of cosolvent on the supercritical fluid
extraction

of bioactive compounds from Helianthus annuus L, J. Supercrit. Fluids 41

(2007) 4349.
[5] F. Macas, A. Torres, T. Galindo, R. Varela, J. Alvarez, J.M.G.
Molinillo, Bioactive terpenoids from sunflower leaves cv Peredovick,
Phytochemistry
61 (2002) 687693.