The document discusses supercritical fluid extraction of borage (Borago officinalis L.) seeds using pure CO2 and mixtures of CO2 with caprylic acid methyl ester. It finds that increasing the CO2 pressure from 100 to 350 bar increased the borage seed extract yield from 0.14% to 24.29% by weight. Adding caprylic acid methyl ester as an entrainer further increased the extract yields, with the highest solubility of the entrainer in CO2 determined to be approximately 1 g of ester in 1 g of CO2 at pressures of 100 and 300 bar. The addition of the entrainer increased extract yields up to 47.8 times at 100 bar and 2

1. Journal of Supercritical Fluids 22 (2002) 211–219
Supercritical fluid extraction of borage (Borago officinalis
L.) seeds with pure CO2 and its mixture with caprylic acid
methyl ester
Egidijus Dauksˇas a
, Petras Rimantas Venskutonis a,
*, Bjo¨rn Sivik b
a
Department of Food Technology, Kaunas Uni6ersity of Technology, Rad6ile; nu¸ pl. 19, Kaunas 3028, Lithuania
b
Food Technology, Chemical Centre, Uni6ersity of Lund, PO Box 124, S-221 00, Lund, Sweden
Received 21 September 2000; received in revised form 22 May 2001; accepted 17 August 2001
Abstract
The influence of different pressures of CO2 and the addition of caprylic acid methyl ester as an entrainer was
studied for the extraction process of borage seed. The increase of CO2 pressure from 100 to 350 bar resulted in the
increase in extract yield from 0.14 to 24.29% (w/w) while the changes in the extract composition were not so
considerable. The highest solubility of pure caprylic acid methyl ester in dense CO2 was determined at 100 and 300
bar (approximately 1 g of ester in 1 g of CO2). The addition of this entrainer increased the yield of pure extract up
to 47.8 times at 100 bar, 2.4 times at 200 and 300 bar. Due to the high solubility of caprylic acid methyl ester at the
lower (100 bar) pressure it is easy to separate the entrainer, which constituted only 4.22% of the total borage seed
extract. © 2002 Elsevier Science B.V. All rights reserved.
Keywords: Borage; Caprylic acid methyl ester; Entrainer; Supercritical CO2
www.elsevier.com/locate/supflu
1. Introduction
Borage (Borago officinalis L.) is an annual or
biannual plant that has a hollow cylindrical stalk
and can reach a height of 80 cm [1]. It is native to
Europe, Asia Minor and North Africa and has
been used for centuries as a culinary herb of
purported medical value. The oil of borage seed
contains g-linolenic acid (18:3v6) and occasion-
ally also stearidonic (18:4v3) acid. Both acids
have a C6-double bond and consequently the 6-
desaturase step is avoided in the biosynthesis of
these acids in man when they are present in food
or food supplement as a ‘medical oil’ [1–5].
Oil from plant seeds consists mainly of triglyce-
rides consisting of C16–C20 fatty acids. Triglyce-
rides are fairly soluble in SC-CO2, but far more so
in n-alkanes, like propane. When appropriate sol-
vents for triglycerides are used, the rate of extrac-
tion can be enhanced at moderate conditions [6].
Fatty acids can be purified by transesterification
with methanol to methyl esters, which can be
hydrolyzed to free fatty acids (FFA) after purifi-
* Corresponding author. Tel.: +370-7-756426; fax: +370-
7-756647.
E-mail address: rimas.venskutonis@ctf.ktu.lt (P.R. Vensku-
tonis).
0896-8446/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S0896-8446(01)00116-4

2. E. Dauksˇas et al. / J. of Supercritical Fluids 22 (2002) 211–219212
cation. To concentrate fatty acids of different
chain length and degree of unsaturation they have
to be removed from glycerol and be processed
either as FFA or as their methyl or ethyl esters
[7].
Fatty acid composition is roughly the same as
that in oils extracted with hexane [8]. Usually, the
content of FFA and of non-saponifiable material,
and also the peroxide value, is of the same order
of magnitude as in the hexane extracts, whereas
the tocopherol content can be higher in the SC-
CO2 extracts [8]. The odor of the extracts is
usually mild and the taste less harsh than that of
crude oils isolated by the traditional process [8].
Lecithin is not soluble in SC-CO2, therefore, the
oil extracted with CO2 contains ten times less
phosphorus than that present in crude oils [8].
The amount of phosphorus in the oil extracted
from oats was considerably increased when oats
were saturated with ethanol before SC-CO2 ex-
traction [9]. Analysis of the extracts from canola
seeds extracted with SC-CO2/ethanol also showed
the presence of phospholipids and long chain fatty
acids (C20:0, C22:0, C22:1), which were not ex-
tracted with pure SC-CO2 [10].
The solubility of lipids can be greatly enhanced
by adding an entrainer to the SC-CO2. In general,
an entrainer enhances solvent power and increases
the effect of pressure and temperature on the fluid
solvent power. By using the optimal co-solvents it
is also possible to improve the effectiveness of the
oil isolation and separation of the desirable com-
ponents from complex matrices [11]. The effec-
tiveness of the separation of oils from co-solvents
and other substances present in the raw material
depends on their differences in volatility and/or
polarity and the effect of solvent density on their
solubility in CO2 [8]. Food grade co-solvents usu-
ally remain in the extracted product and can be
used to obtain required concentrations of the
extracts. For this reason, ethanol is the most
preferred co-solvent for food related applications.
Non-alcohol organic solvent entrainers may have
very limited applications for special industrial
products and processes [12].
Caprylic acid is present in various fat contain-
ing products as a natural component. For in-
stance, it constitutes (GC area%) 2.6–7.3% of
babassu oil, 4.6–9.4% of coconut oil and 2.4–
6.2% of palm kernel oil [13]. The amount of
caprylic acid in milk fat or butter varies in the
range of 0.5–2%. The solubility of tricaprylin was
0.13 g l−1
CO2 [14]. The fatty acid esters are
soluble to a greater extent in dense CO2 than FFA
and are thus preferentially extracted [7].
Borage seed oil was also extracted by supercrit-
ical CO2 and it was found that in the optimal
working conditions the yield was comparable with
that obtained by the conventional extraction with
hexane [15]. The objectives of the present study
were to determine the effect of CO2 pressure on
the extract yield from borage seeds and to exam-
ine the possibilities of the yield increase by using
caprylic acid methyl ester as extraction etrainer.
2. Materials and methods
2.1. Materials
The seeds of borage (B. officinalis L.) were
collected from the experimental garden of Lithua-
nian Institute of Horticulture in 1998. The seeds
were harvested manually, dried at 30 °C in a
ventilated drying oven (Vasara, Utena, Lithuania)
and stored in paper bags at ambient temperature
protected from light until further analysis. The
samples were ground prior extraction by Knifetec
1095 Sample Mill (Tecator AB, Ho¨gana¨s, Swe-
den) machine for 20 s.
The following chemicals were used for the ex-
periments: carbon dioxide (99.99%) from Aga Gas
(Stockholm, Sweden), diethyl ether and hexane
(99%) from Merck (Darmstad, Germany),
caprylic acid methyl ester (methyl octanoate) from
ICN (Costa Mesa, CA, USA), boron trifluoride
(12% in methanol) and sodium methoxide (0.5 M
in methanol) from Acros (NJ, USA), Celite 545,
20–45 mm from Kebo lab (Prolabo, France).
2.2. Extraction apparatus and methodology
A schematic diagram of the experimental ap-
paratus used in this study is shown in Fig. 1. A
Milroyal B–C pump (Dosapro Milton Roy,
Pont–Saint–Pierre, France) was used for the ex-

3. E. Dauksˇas et al. / J. of Supercritical Fluids 22 (2002) 211–219 213
Fig. 1. Schematic drawing of the supercritical extraction equip-
ment. (1) Gas tube, (2) shut-off valve, (3) gas filter, (4) ethanol
bath, −22 °C, (5) pump, (6) safety valve, (7) pressure gauge,
(8) shut-off valve, (9) extractor, (10) water bath, (11) micro
metering valve, (12) test tube, (13) cooling bath, (14) flowmeter
(15) entrainer supply pump.
in the test tube held at −5 °C cold bath. The
first experiment was performed by passing 200 g
(100 l) of CO2 through 10 g of borage seeds. The
solvent was evaporated from the extract at atmo-
spheric pressure and measurements were made
after passing 1–2 g CO2 per 1 g borage seeds Fig.
2. The precision balances (Mettler AE 163, read-
ability 0.01 mg, Mettler Instrumente AG, Switzer-
land) were used to weigh the extracted oil. Two
replicates were analyzed for every sample and the
mean value calculated.
The solubility of caprylic acid methyl ester was
determined by extracting 10 g of caprylic acid
methyl ester on 20 g Celite in 47 ml extractor with
10 l (20 g) of CO2 at 100, 200 and 300 bar
pressure, 40 °C temperature and CO2 flow rate of
0.5 l min−1
.
2.3. GC conditions
The oil (150 ml) was first transesterified with
boron trifluoride-methanol and 0.5 M methanolic
sodium hydroxide, and then the fatty acid methyl
traction. The seed samples were covered by glass
wool on the bottom and top of the cell. Extrac-
tions were performed at 40 °C temperature, and
100–350 bar pressure with or without caprylic
acid methyl ester addition as a co-solvent, 0.5–2%
(w/w) of CO2. Caprylic acid methyl ester was
dosed by an extra pump 15 (Fig. 1). Ten grams of
plant material were extracted in a 47 ml capacity
vessel at CO2 flow rate of 0.5 l min−1
(measured
at atmospheric pressure) by collecting the extract
Fig. 2. The yield of borage seed oil as a function of the amount of CO2 at different pressures and 40 °C.

4. E. Dauksˇas et al. / J. of Supercritical Fluids 22 (2002) 211–219214
esters (FAME’s) were extracted into hexane as
described in AOAC method 969.33 [16].
FAMEs were analyzed on a Varian 3400 capil-
lary gas chromatograph equipped with a flame
ionization detector connected to a Vista 420 inte-
grator (Varian Associates, Walnut Creek, CA)
and a fused silica capillary column, Supelcowax™
10; 60 m, 0.32 mm id, 0.50 mm film thickness
(Supelco Inc., Bellefonte, PA). The oven tempera-
ture was held at 180 °C for 8 min, then increased
to 225 °C at 10 °C min−1
and held for 28 min.
The temperature of the on-column injector was
raised from 180 to 250 °C at 100 °C min−1
and
kept at 250 °C for 30 min. The temperature of
detector was 250 °C. Helium was used as a car-
rier gas at a flow rate of 4 ml min−1
. For the
determination of caprylic acid methyl ester in the
samples oven temperature program was started at
50 °C (5 min hold) and then raised to 225 °C (25
min hold) at 10 °C min−1
.
FAMEs were identified by the comparison of
their retention times with those of a reference
solution chromatographed at identical GC condi-
tions. Two replicate GC analyses were performed
and the results were expressed in GC area% as a
mean value. The amount of caprylic acid methyl
ester in the extracts was also expressed in GC
area% as its peak eluted in the chromatograms
together with FAMEs of borage oil. It was as-
sumed that GC area% of a caprylic acid methyl
ester peak is proportional to its content in the
extract. Consequently the sum of GC area% of
borage oil FAMEs peaks represents the content of
borage oil in the extracts.
3. Results and discussion
3.1. Effect of CO2 pressure on borage seed
extraction
The extraction curves (in g extract per g seeds)
obtained during the first experiment show that
CO2 at 100 bar dissolves only negligible amount
of borage substances (approximately 0.14% (w/w)
after passing 20 g CO2 g seeds). When the extrac-
tion pressure was raised up to 150 and 200 bar,
the total yield of the extract increased to 5.59 and
15.15% (w/w), respectively. The highest extract
yield and the fastest extraction rate were achieved
after increasing the pressure above 250 bar. The
extraction rate increased slightly with an increase
of pressure from 250 to 300 and 350 bar, however,
the final yields were 21.89, 21.59 and 24.29%,
respectively, which were quite comparable for all
these pressures. The appearance of the extracts
obtained at different pressures was very similar
and can be characterized as a yellow oil-like trans-
parent liquid. However, some changes in the in-
tensity of the color (darker at higher pressures)
were observed.
Fatty acid composition of the extracts obtained
at different pressures is presented in Table 1. The
major fatty acids were C16:0, C18:1, C18:2,
C18:3v6. C16:0 content decrease from 14.00 to
10.04% occurred upon increasing the pressure
from 100 to 350 bar. C18:3v6 content increased
from 16.21 to 20.06% by increasing the pressure
from 100 bar up to 200 bar, however, it slightly
decreased (to 18.51%) when the fluid was further
pressurized up to 350 bar. The content of longer
chain unsaturated fatty acids, C20:1, C20:2, C22:1
and C24:1 increased with increasing pressure.
Their percentage in the oil obtained at 250–350
bar was on the average 1.5–3 times higher than in
the oil extracted at 100–150 bar.
3.2. Effect of caprylic acid methyl ester entrainer
on borage seed extraction
Fig. 3 illustrates the dependency of pure
caprylic acid methyl ester solubility on the
amount of CO2 at different pressures (100, 200
and 300 bar). The extraction kinetics curves show
that the solubility of caprylic acid methyl ester
was higher at 100 bar and 300 bar than that
obtained at 200 bar.
Caprylic acid methyl ester was used as an en-
trainer in further experiments, which were carried
out at three different extraction pressures: 100,
200 and 300 bar. Caprylic acid methyl ester was
dosed in to the extraction system to constitute 0.5,
1 and 2% (w/w) of the entrainer based on the
amount of main solvent, CO2. The extracts ob-
tained in the test tube were subjected to the
transesterification procedure and the amount of

5. E. Dauksˇas et al. / J. of Supercritical Fluids 22 (2002) 211–219 215
Table 1
Fatty acid composition (GC area%) of borage seeds oil extracted with SC-CO2 at different pressures
Pressure (bar)Fatty acids
100 150 200 250 300 350
12.84 10.8216:0 10.6814.00 10.39 10.04
0.3916:1 0.760.46 0.44 0.65 0.32
0.47 0.60 0.460.50 0.5516:2 0.38
1.5318:0 1.80 1.66 1.55 1.06 1.03
23.8918:1 22.94 21.57 22.07 22.49 22.38
34.92 35.75 35.0734.53 34.7618:2 34.18
17.95 20.06 19.29 19.20 18.5118:3v6 16.21
nd nd nd1.00 nd18:3v3 nd
0.2618:4v3 0.77 nd nd nd nd
0.48 nd 0.3320:0 0.350.71 0.36
3.79 4.30 4.783.86 4.7720:1 5.08
nd20:2 nd nd nd nd 0.29
1.7922:1 2.32 2.80 3.25 3.44 4.05
1.31 1.69 2.091.01 2.3524:1 3.07
99.75Total 100.00 100.00 100.00 100.00 99.72
nd; not detected.
Fig. 3. The yield of caprylic acid methyl ester as a function of the amount of CO2 at different pressures and 40 °C.
caprylic acid methyl ester in the extracts was
determined by GC (Table 2). When 0.5% of en-
trainer was added for the extraction at 100 bar the
content of caprylic acid methyl ester in the extract
constituted 51.12% of the total FAMEs. After
increasing the pressure to 200 and 300 bar the

6. E. Dauksˇas et al. / J. of Supercritical Fluids 22 (2002) 211–219216
Table 2
Caprylic acid methyl ester content (GC area %) of SC-CO2
extracts obtained at different pressures
Amount of caprylic acid methyl esterPressure (bar)
added into CO2, % (w/w)
1 20.5
72.96100 79.6951.12
20.83 39.76200 11.87
14.32 23.37300 4.22
sure increased the total yield of the extract and
consequently the percentage of the entrainer in
the extract was reduced.
The amounts of pure borage oil extracted with
entrainer addition were calculated by subtracting
the amount of caprylic acid methyl ester in the
extracts (determined by GC) and the graphs rep-
resenting the effect of entrainer on the yield were
plotted. The influence of different amounts of
entrainer on the yield at 100 bar is illustrated in
Fig. 4. The yield of pure borage seed oil after
passing 20 g CO2/l g seed at this pressure was very
low, approximately 0.14%. When 0.5, 1 and 2% of
the entrainer were added, the final extract yield
was 2.93, 3.06 and 6.70%, respectively. Compared
with the very small yield obtained at 100 bar with
pure CO2 the use of entrainer was very effective
for solubilizing more oil.
The use of entrainer was also very effective at
higher extraction pressures. The addition of 0.5, l,
and 2% of entrainer at 200 bar Fig. 5 enabled to
increase the yield to 20.69, 30.04, and 36.58%
(w/w), respectively, compared with 15.15% ob-
tained with pure CO2 at the same extraction con-
ditions. When the extraction was carried out at
300 bar pressure the curves acquired a somewhat
percentage of the entrainer residue in the extracts
was considerably reduced due to the significant
increase in the total extract yield and constituted
only 11.87 and 4.22%, respectively. When the
dosage of the entrainer was increased to 1 and
2%, its content in the extract was also higher. For
instance, the addition of 1% of caprylic acid
methyl ester increased its content in the extract to
72.96% at 100 bar, 20.83% at 200 bar and 14.32%
at 300 bar; when 2% entrainer were used, the
extracts obtained at 100, 200 and 300 bar con-
tained 79.69, 39.76 and 23.37% of the entrainer,
respectively. In any case, the increase of the pres-
Fig. 4. The influence of entrainer on the yield of borage seed oil at 100 bar pressure.

7. E. Dauksˇas et al. / J. of Supercritical Fluids 22 (2002) 211–219 217
Fig. 5. The influence of entrainer on the yield of borage seed oil at 200 bar pressure.
Fig. 6. The influence of entrainer on the yield of borage seed oil at 300 bar pressure.
logarithmic function shape Fig. 6. The final yield
obtained with 0.5, 1, and 2% of entrainer was
30.72, 38.18, and 51.54% (w/w), respectively, com-
pared with 21.59% obtained with pure CO2 at the
same extraction conditions. The results obtained
clearly demonstrate that both the pressure and the
amount of the entrainer have a substantial effect
in increasing the extract yield.

8. E. Dauksˇas et al. / J. of Supercritical Fluids 22 (2002) 211–219218
The fatty acid composition of borage seed oil
in the extracts obtained with different levels of
caprylic acid methyl ester as entrainer is pre-
sented in Table 3. Some differences in the fatty
acid composition of various extracts are appar-
ent. The content of palmitic acid decreases with
an increase of pressure independent of the added
amount of entrainer. The percentage range of the
major acids were: oleic was from 18.40 to 25.63%
(decreased when the pressure was raised and 0.5
and 2% of entrainer used), linoleic from 32.64 to
37.72% (decreased with pressure 1% of en-
trainer), and g-linolenic from 14.74 to 20.66%
(increased with pressure at 1 and 2% entrainer).
The content of some other fatty acids also varied
in a similar range, however, it is difficult to find
clear effect of the entrainer or pressure on these
changes within the scope of the present study.
The decrease in the palmitic acid content with
pressure was quite similar in case of pure CO2
(Table 1) and in case of application of various
doses of ester (Table 3). The opposite tendency
can be observed with fatty acids consisting of
more than 18 carbon atoms.
4. Conclusions
The yield of the extract from borage seed in-
creases when CO2 pressure increases from 100 to
350 bar. Some effect of the extraction pressure on
the fatty acid composition of the extracts was
determined. The caprylic acid methyl ester is solu-
ble in dense carbon dioxide (almost 1 g in 1 g of
CO2) and can be easily removed from the borage
seed extract at lower pressure. This compound,
when used as an extraction entrainer results in an
increase in the amount of the extract from 0.14 to
6.7% at 100 bar, from 15.15 to 36.58% at 200 bar
and from 21.59 to 51.54% at 300 bar and 40 °C.
Acknowledgements
The authors wish to thank Lithuanian Insti-
tute of Horticulture for providing plant material,
Johnson Foundation (Royal Institute of Technol-
ogy, Sweden) for the financial support to carry
out experiments and Lithuanian State Founda-
tion of Science and Studies for the partial aid to
conduct this study.
Table 3
The influence of entrainer amount on the fatty acid composition (GC area%) of borage seed oil extracted at different pressures
2% of entrainer1% of entrainer0.5% of entrainerCompounds
100300200100 300300200100 200
10.4212.4615.2810.9711.8716:0 15.9611.2511.9515.83
0.39 nd16:1 1.280.61 0.75 0.590.32 0.16 0.96
1.91 1.2718:0 2.171.92 1.53 1.431.14 1.95 1.66
21.2322.2724.8925.6324.5618:1 24.0518.4021.0921.53
37.72 35.24 35.55 32.64 33.03 34.01 34.5818:2 35.7135.52
18:3v6 18.7416.3716.0715.8715.6914.7419.4420.6617.53
1.222.181.09nd 1.091.44ndnd1.2918:3v3
20:0 nd0.63 0.24 0.58 0.67 0.53nd nd 0.77
2.73 4.29 4.71 2.9520:1 4.57 4.99 2.50 4.62 4.88
20:2 nd nd 0.39 nd 0.21 0.45 nd 0.65 0.80
2.43 3.743.501.233.772.891.22 1.3322:1 2.64
1.19 1.42 1.65 0.9124:1 1.96 2.39 0.80 1.95 1.96
100.00 99.01Total 98.31 100.00 99.59 99.32 100.00100.00100.00
nd; not detected.

9. E. Dauksˇas et al. / J. of Supercritical Fluids 22 (2002) 211–219 219
References
[1] A. Medrano, T.A. Masound, M.C. Martinez, Mineral
and proximate composition of borage, J. Food Comp.
Anal. 5 (1992) 313.
[2] A. Whipkey, J.E. Simon, J. Janick, In Vivo and in vitro
lipid accumulation in Borago officinalis L, J. Am. Oil
Chem. Soc. 65 (1988) 979.
[3] P. Sewo´n, E. Tystja¨rvi, Stearidonic and g-linolenic acid
contents of common borage leaves, Phytochemistry 33
(1993) 1029.
[4] A. Sensidoni, G. Bortolussi, C. Orlando, P. Fantosi, G.
Lognay, M. Paquot, Composition and oxidative stability
of borage (Borago officinalis L.) and borage-virgin olive
oil blends, Lebensm-Wiss u-Technol. 28 (1995) 343.
[5] M.S.K.S. Rahmatullah, V.K.S. Shukla, K.D. Mukherjee,
Enrichment of g-linolenic acid from evening primrose oil
and borage oil via lipase-catalysed hydrolysis, J. Am. Oil
Chem. Soc. 71 (60) (1994) 569.
[6] G. Brunner, Gas Extraction, Springer, New York, 1994.
[7] A. Staby, J. Mollerup, Solubility of fish oil fatty acid
ethyl esters in sub- and supercritical carbon dioxide, J.
Am. Oil Chem. Soc. 70 (1993) 583.
[8] E. Stahl, K.W. Quirin, D. Gerard, Dense Gases for
Extraction and Refining, Springer, Berlin, 1988.
[9] S.M. Fors, C.E. Eriksson, Characterization of oils ex-
tracted from canola seeds with supercritical CO2, Lebens-
Wiss u Technol. 23 (1990) 390.
[10] F. Temelli, Extraction of triglycerides and phospholipids
from canola seeds with supercritical CO2 and ethanol, J.
Food Sci. 57 (1992) 440.
[11] S. Peter, Supercritical fractionation of lipids, in: J.W.
King, G.R. List (Eds.), Supercritical Fluid Technology in
Oil and Lipid Chemistry, AOCS press, Champaign, IL,
1996, p. 65.
[12] D.A. Moyler, B. Road, CO2 extraction of essential oils.
Part III, pimento berry, coriander and celery seed oil, in:
G. Charalambous (Ed.), Flavors and Off-Flavors, El-
sevier, Amsterdam, 1989, p. 263.
[13] Codex Alimentarius Commission, Joint FAO/WHO
Foods Standard Program,XXII Session, Geneva 23–28
June, 1997, pp. 51.
[14] H. Hammam, Solubilities of pure lipids in supercritical
carbon dioxide, J. Supercrit. Fluids 5 (1992) 101.
[15] A.D. Andujar, A.M. Gomez, E.M. de la Ossa, Character-
ization of borage seed oil extracted by supercritical car-
bon dioxide, Grasas Aceites 50 (1999) 283.
[16] AOAC. Official Methods of Analysis, 15th ed. of the
Association of Official Analytical Chemists, Washington,
DC, 1990.