food and bioproducts processing 9 4 ( 2 0 1 5 ) 365–374

Contents lists available at ScienceDirect

Food and Bioproducts Processing

journal homepage: www.elsevier.com/locate/fbp

Study of supercritical extraction from Brazilian

cherry seeds (Eugenia uniflora L.) with bioactive
compounds

Débora Nascimento e Santos, Larissa Lima de Souza, Nilson José Ferreira,

Alessandra Lopes de Oliveira ∗
Food Engineering Department, Faculty of Animal Science, University of São Paulo, Av. Duque de Caxias Norte, 225,
Jardim Elite, CEP 13635-900, Pirassununga, SP, Brazil

a b s t r a c t

In this study, we obtained extracts from Brazilian cherry (or Pitanga) seeds using supercritical CO2 (SC-CO2 ) and
analyzed the volatiles profile. Seeds were placed in contact with the SC-CO2 under different conditions of pressure
(P) and temperature (T). The yield, the refractive index, the presence of terpenes by thin layer chromatography (TLC)
and the volatiles profile by gas chromatography with mass spectrometry (GC–MS) were determined for all the extracts.
With respect to the extract yield (0.16–0.48 g/100 g), surface response analysis revealed that the greatest yield was
obtained at high P and T did not affect the process, which indicated that low temperatures can be used to protect
the thermosensitive substances. The presence of terpenes with antioxidant activity was detected via TLC analysis,
and the GC–MS results indicated ␥-elemene and germacrone as the major components and the concentration of the
sesquiterpenes in the extract was influenced by the interaction between P and T.
© 2014 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Keywords: Eugenia uniflora L.; GC–MS; Sesquiterpenes; Response surface analysis

1. Introduction ideal solvent for the extraction of natural products, because

In recent years, interest has been an increasing in the recovery
of bioactive compounds from natural sources for the develop-
ment of functional foods. Supercritical fluid extraction (SFE)
has been recognized as an increasingly applied in the food,
pharmaceutical and cosmetics areas; it provides higher selec-
tivity and shorter extraction times and does not involve the
use of toxic organic solvents. Because of these and other
advantages, SFE has been used in various industries (Marsal
et al., 2000; Serra et al., 2010). A supercritical fluid is defined
as the state in which liquid and gas are indistinguishable.
Because of its low viscosity and high diffusion capacity, super-
critical fluids exhibit better transport properties than liquids,
which allows rapid extraction can be accomplished (Berna
et al., 2000; Andreo and Jorge, 2006). Carbon dioxide is an
it exhibits low toxicity and is non-explosive, readily available
and easy to remove from the extracted solutes. The use of CO2
makes part of this extraction method free of organic solvents
and the method is therefore considered environmentally
friendly. Additionally, because the critical temperature of CO2
is 31 ◦ C, thermosensitive natural compounds can be recovered
with minimal possibility of damage or decomposition (Salgin,
2007; Danh et al., 2009; Laroze et al., 2010).
The extracts obtained by SFE using CO2 are generally
recognized as safe (GRAS) for use in food production, such
as the extraction of oil from a large number of materi-
als whose extracts have shown strong potential for future
applications in drugs and food (Zhang et al., 2010). The extrac-
tion with supercritical CO2 is a powerful tool in research
on fragrances and essential oils, especially because the

∗
Corresponding author. Tel.: +55 19 3565 4268; fax: +55 19 3565 4284.
E-mail address: alelopes@usp.br (A.L.d. Oliveira).
Received 19 December 2013; Received in revised form 20 March 2014; Accepted 20 April 2014
Available online 29 April 2014
http://dx.doi.org/10.1016/j.fbp.2014.04.005
0960-3085/© 2014 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
366 food and bioproducts processing 9 4 ( 2 0 1 5 ) 365–374
isolation of volatile compounds using traditional methods,
such as steam distillation or dynamic headspace sampling
produces fragrance extracts that may not reflect the real
flavor properties of the natural material. With extraction
conditions that use mild temperatures, supercritical fluid
extraction is the most appropriate method to isolate essen-
tial oils from spices, flowers, herbs, leaves, seeds and roots.
The extracts are generally regarded as superior to sensory
concentrates produced using traditional techniques (Marsili,
2002).
The Pitanga fruit or Brazilian cherry (Eugenia uniflora L.)
belongs to the botanical family Myrtaceae. It is a native
plant to Brazil and the country has been estimated to
be the world’s largest producer, although it is also found
in Northern Argentina and Uruguay (Bezerra et al., 2004;
EMBRAPA, 2006). The nutritional value of Pitanga is rele-
vant, because it high concentrations of calcium, phosphorus,
pro-vitamin A, vitamin C and phenolic compounds (antho-
cyanins and carotenoids), the latter of which are known
for their antioxidant activity (Mélo et al., 1999; Silva, 2006;
Lopes Filho et al., 2008). The leaves and fruits of Eugenia uni-
flora L. in the form of infusions have been widely used in
Brazilian folk medicine with several effects, such as excit-
ing, febrifuges, aromatic, antidysenteric, antihypertensive and
antirheumatic and its alcoholic extract is used to treat bron-
chitis, coughs, fevers, anxiety and verminosis (Auricchio et al.,
2007). The hydroalcoholic leaves extract decreased levels of
the xanthine oxidase enzyme, linked to the onset of gout,
and also presented features of vasorelaxant, antioxidant,
antidiarrheal, hypotriglyceridemic, hypoglycemic and with
antibacterial property (Ogunwande et al., 2005; Amorim et al.,
2009).
The work by Oliveira et al. (2006) shows that the Brazil-
ian cherry fruit has monoterpenes similar to those present in
the leaves, whose bioactivity has been scientifically proven,
explaining the similarity of the effects seen in both extracts.
Considering the fruit particularity, the same research group
studied the process optimization to obtain cherry extracts by
the use of supercritical CO2 (Malaman et al., 2011). Sensory
tests showed the extracts were rich in volatile compounds,
leading to the perception of greater intensity of the fruit
aroma; and according to the literature, these compounds are
also present in the leaves (Pino et al., 2003; Oliveira et al.,
2009).
On the seeds, there are few studies reporting its composi-
tion. Bagetti et al. (2009) evaluated the nutritional composition
and antioxidant activity of these seeds and the results
revealed carbohydrates as the most abundant, composed pri-
marily of insoluble fiber. The protein fraction is very small.
Among the fatty acids present, 45–47% are polyunsaturated,
desirable for the human diet have an effect on reducing the
incidence of cardiovascular disease. In another study car-
ried out by Luzia et al. (2010), it was obtained ethanolic
extracts from Pitanga seeds showed higher antioxidant poten-
tial with a high concentration of phenolic compounds. Among
the phenolic compounds found in Pitanga seeds ethanolic
extracts obtained by pressurized fluid extraction are present
flavonol and conjugate acids with sugars (Oliveira et al.,
2014).
In this study, supercritical extract was obtained from the
Pitanga seeds and with this objective was to analyze the
behavior of pressure and temperature in the yield and the
volatile compounds extraction when these conditions was
applied.
2. Materials and methods
2.1. Materials
Ripe native fruit were collected on the Campus of the Faculty
of Animal Science and Food Engineering (FZEA), University of
São Paulo (USP), Pirassununga (21◦ 49 46 S 47◦ 25 33 W), Brazil.
In the Laboratory of High Pressure Technology and Natural
Products (LTAPPN), the seeds were selected, washed to remove
traces of pulp, and dried in an oven with air circulation at 38 ◦ C
for 54 h. The dried seeds were peeled and crushed in ham-
mer mill type (Buhlermiag, Uzwil, Switzerland) and then kept
frozen at −20 ◦ C.
Petroleum ether, ethanol (EtOH), chloroform and sodium
carbonate were obtained from Synth (Diadema, Brazil). Gallic
acid and the homologous series of alkanes C10–C40 were pur-
chased from Fluka (Sant Louis, Unites States). R-(+)-limonene,
radical 2,2-diphenyl-1-picrylhydrazyl (DPPH) and anisalde-
hyde were purchased from Sigma Aldrich (Sant Louis, Unites
States). Chromatographic plate Silica gel 60 F254 and toluene
were obtained from Merck GE. We also used ethyl acetate
(Mallinckrodt Chemicals, Dublin, Ireland) and methanol (JT
Baker, Dublin, Ireland).
The moisture content of the seeds after dryer was deter-
mined by gravimetric method which determines the mass loss
from the sample by drying to constant weight (AOAC, 1995).
The ether extract was obtained by the hot extracting lipids
method using Soxhlet with petroleum ether as organic sol-
vent (IAL, 2008). The crushed seeds (200 g) granulometry was
determined in a set of six Tyler standard sieve series with stir-
ring for 15 min (ASAE, 2008 – S319.4). For the determination of
apparent density, we used the fixed bed of 300 cm3 (extractor)
as standard container volume. For 10 times it was completed
this volume with the ground Pitanga seeds and after removal
of the extractor, the mass was weighed. The apparent density
was obtained by the ratio between the average mass of seeds
and volume of fixed bed.
2.2. Supercritical fluid extraction (SFE)
In the extraction procedure it was promoted the contact
between the crushed seeds, which were packed in fixed bed
extractor 300 cm3 , with supercritical CO2 to pre-set pressure
(P) and temperature (T) using a central composite design (CCD).
The pressure was performed by a high pressure pump (Eldex,
S-100-AA, United States), while the temperature was main-
tained by operation of a thermostatic bath, in which the
extractor has been packaged. The extract, separated from the
gas was collected in a flask and CO2 flow rate (2 g CO2 /min),
was measured in a rotameter (LAO, G-1, Osasco, Brazil) under
ambient conditions (0.94 bar and 25 ◦ C). At the end of the pro-
cess, the extracts were weighed and related to the mass of
dry seeds packed in the extractor to calculate the overall yield
(Malaman et al., 2011).
The operating conditions of the experimental procedure
were made according with a CCD with P and T as independent
variables (Table 1) totaling 11 tests (Table 2), which were per-
formed at random. For all tests, it was standardized a time of
16 h of static equilibrium and dynamic extraction period of 6 h.
The static extraction time has been applied in order to ensure
that the mechanical and thermal equilibrium is guaranteed in
the employed operating conditions.
 food and bioproducts processing 9 4 ( 2 0 1 5 ) 365–374
Table 1 – Levels of independent variables (P and T) for the central composite design.
Variables
−˛ −1
P (kgf/cm2 ) 83 110
T (◦ C) 31 35
˛ = ±1.41.
2.3. SFE of Pitanga seeds with EtOH and extracts
purification
In the SFE-EtOH procedure, EtOH was used as the co-solvent.
Seeds were prepared as described above and mixed with EtOH
(35 mL/72 g of crushed seeds, approximately 1:2 mL of EtOH:g
seeds) before the extraction. Next, the seeds were placed in
the extractor and the extraction was performed with a super-
critical pressure and temperature of 175 kgf/cm2 and 45 ◦ C,
respectively. These conditions were those of the CCD central
point and produced the best per cent yield. Similar to the
supercritical extractions performed with no modifier polari-
ties, this extraction was also conducted for a standard period
of 16 h to establish static equilibrium and the dynamic extrac-
tion of period was 6 h.
Purification of the extract from SFE-EtOH was performed
using the method described by Lhuillier et al. (2007) with mod-
ifications (Oliveira et al., 2014).
2.4. Extracts characterization
The refractive index of supercritical extracts from Pitanga
seeds was measured at 20 and 40 ◦ C in refractometer (LAMBDA
2WAJ, ATTO Instruments Co., Tsuen Wan, Hong Kong) accord-
ing to IAL (2008). In view of the presence of terpenes, the
extracts were applied on silica gel Plates 60 F254 (stationary
phase) and eluted with toluene and ethyl acetate (93:7, v:v)
as the mobile phase. After spraying the developer, an acidic
solution of Anisaldehyde, the plates were dried and heated
to visualize the bands that eluted with varied coloration, pre-
pared as described by Wagner and Bladt (2009). In preliminary
analysis of antioxidant activity in these extracts it was used
as mobile phase a solution of toluene, ethyl acetate (95:5, v:v),
and a methanol solution of 0.05 g/100 g of 2.2 radical-diphenyl-
1-picrylhydrazyl (DPPH• ) as developer, sprinkled on the plate
after elution and drying. Those spots that were yellow on
purple background showed indicative of redox reaction, indi-
cating components with antioxidant activity.
The volatile composition of the extracts was analyzed
by gas chromatography coupled with mass spectrometry
(GC–MS) (QP 2010 Plus, Shimadzu, Tokyo, Japan) with autosam-
pler (AOC-5000, SWI, Tokyo, Japan). In this analysis was used
a capillary column Rtx® -5MS (30 m × 0.25 mm id × 0.25 ␮m df,
5% diphenyl/95% dimethylpolysiloxane) (Restek, Bellefonte,
United States) and helium as carrier gas to an internal pres-
sure of 60 psi. The extract diluted in methanol (400 mg/L) 1 ␮L
was injected into a split ratio of 1:20. The injector tempera-
ture was 220 ◦ C and the interface between the ion source and
the column was 250 ◦ C. The oven temperature was regulated
so as to obtain a linear gradient from 60 to 246 ◦ C at 3 ◦ C/min
(Adams, 2009). The source ionization energy was 70 eV, and
the mass scan ranged from 50 to 500 m/z.
367
Levels
0 +1 +˛
175 240 267
45 55 59
The spectra identification was performed using software
GC–MS solutions v. 2.5 (NIST 08 and NIST 08s) database
libraries. Standard alkane solution (C10–C40) prepared in
methanol was applied in the GC–MS under the same con-
ditions for extracts analysis to calculate the Kovats index
(KI) and Van der Dool and Kratz (AI) (Adams, 2009). The
values found were compared with the literature data, for
the same analysis conditions. For the quantification of the
major sesquiterpenes, we used the external standardiza-
tion method, using as external standard the monoterpene
R-(+)-limonene (y = 48221.1229 + 21191.0373x; R2 = 0.9957) in
different concentrations in the same conditions for GC–MS
analysis of the extracts.
3. Results and discussion
3.1. SFE
Dried and crushed Brazilian cherry seeds had moisture con-
tent of 12.73 ± 0.17. Since drying performed this value is
below to found in the literature, around 58 g/100 g (Bagetti
et al., 2009). The average value of ethereal extract obtained
was 0.52 ± 0.07 g/100 g, similar to that found in other studies
(Bagetti et al., 2009; Luzia et al., 2010). The average appar-
ent density (a ) measured was 0.482 g/cm3 . The mass retained
in the six series Tyler sieves suitable for particle size anal-
ysis resulted in the calculated average particle diameter of
0.48 mm. Oliveira (2010) considers an average particle diam-
eter ideal for extraction between 0.25 and 2 mm, which covers
the results found in this study. This parameter impact on the
internal resistance to mass transfer, where often to increase
the extraction yield, it is possible to reduce the particle size,
in order to increase the contact area between the solid and
solvent, falling thus the distance that the solute covers within
the porous particles to the surface. However, reducing the solid
particle size cannot be performed indefinitely, since very small
particles tend to form clusters, and retard the passage of the
fluid through the bed, and cause loss of volatile compounds in
the vegetable matrix rich in essential oils, for example.
Table 2 shows the extraction yield values, the refractive
index data and, the concentration of the major volatile com-
pounds, ␥-elemene and germacrone. For each test or for each
supercritical operational condition of T and P there is a corre-
sponding the density of CO2 calculated by the equation Huang
et al. (1985) with the aid of a computational program (PCO22,
v.1 01-97). When assessing the overall performance by con-
cerning it to the densities of supercritical CO2 , it appears that
the high densities qualified to a higher power supercritical
fluid salvation of the solute and, consequently, higher yields
of these extracts. There is no separately discussing the effect
of density, pressure and temperature of the supercritical CO2 .
For example, the yield obtained in test 8, (CO2 = 674.89 kg/m3 )
was as high as that obtained in test 2 (CO2 = 896.08 kg/m3 ) and
368 food and bioproducts processing 9 4 ( 2 0 1 5 ) 365–374

Volatiles in 100 g of seeds (mg LimEq)

Table 2 – Complete CCD matrix (including three central points) to study P and T in the Yield (Y), the refractive index (RI) and the main sesquiterpenes in the supercritical extracts

Germacrone

0.0047
0.0098
0.0088
0.0052
0.0029
0.0080
0.0059
0.0066
0.0057
0.0064
0.0067
␥-Elemene

0.0048
0.0104
0.0091
0.0076
0.0024
0.0108
0.0063
0.0076
0.0072
0.0084
0.0079
Volatiles in the extracts (mg LimEq/kg)

Germacrone

16.66
21.35
38.35
11.20
18.28
19.16
15.24
14.41
13.00
16.15
14.00
␥-Elemene

17.38
22.67
39.66
16.21
15.18
25.85
16.39
16.73
16.45
21.21
16.62
RI 20 ◦ C

1.5169a
1.5194a
1.5159a
1.5034a
1.5139a
1.5074a
1.5184a
1.5129a
1.5164a
1.5144a
1.5084

Fig. 1 – Study of the independent variables in the yield of

Pitanga seed extracts. (a) Pareto chart of the variables (P
and T) to first order effects analysis. (b) Response surface
model of pressure and temperature in the yield generated
Yield (g/100 g)

by the second order model.

IR with the same letters are not significantly different at 95% confidence limit.
0.28
0.46
0.23
0.47
0.16
0.42
0.39
0.46
0.44
0.40
0.48

this is due to the influence of temperature that in the test 8 is

59 ◦ C, whereas in the test 2 is 35 ◦ C (Table 2).
The yields of the supercritical extracts from Brazilian
CO2 (kg/m3 )

cherry seeds ranged from 0.16 to 0.48 g/100 with a mean of

0.38 g/100 g. Analyzing the influence of the factors studied (P
747.17
896.08
420.89
802.33
242.37
865.54
862.77
674.89
777.57
777.57
777.57

and T) on the extraction yield using supercritical CO2 , the

Pareto diagram (Fig. 1a) shows the analysis of effects, just the
pressure was significant for this response (P < 0.05). The lin-
ear and quadratic models presented in Table 3 show that the
significant coefficients were those related to the P. The inter-
(◦ C)

35
35
55
55
45
45
31
59
45
45
45

action between P and T, showed no significant coefficients

T

in the models. In the analysis of variance (ANOVA) (Table 3)

from Brazilian cherry seeds.
Variables

both models show linear and quadratic significant effects for

at least one variables (Fcalculated > Ftabulated , P < 0.05). The coeffi-
(kgf/cm2 )

cients of determination (R2 ) (0.76 for the linear model and 0.95
110
240
110
240
83
267
175
175
175
175
175

for the quadratic model) suggest that the quadratic model has
a good fit of experimental data showing a good for predicting
Central point.
P

the ability for the yield of extract.

In the response surface analysis (RSA) generated by the
quadratic model (Fig. 1b), the optimal region is that with higher
Essay

values of P were applied for the entire range of T studied. The

9e
10e
11e
1
2
3
4
5
6
7
8

response surface shows the response behavior in function of

 food and bioproducts processing 9 4 ( 2 0 1 5 ) 365–374 369

Table 3 – Analysis of variance for first and second order factorial experimental design for Brazilian cherry seeds extract
yield (g/100 g).
Source of variation Sum square Degree of freedom Mean square Fcalc Ftab
(SS) (DF) (MS)

Linear model (first order) – R = 0.394 + 0.105P

Model (R) 0.057 1 0.057 9.5 6.61
Residue (r) 0.003 5 0.0006
Total 0.060 6
Quadratic model (second order) – R = 0.22 + 0.05P − 0.035P
Model (R) 0.114 2 0.057 152 4.46
Residue (r) 0.003 8 0.000375
Total 0.117 10

Fcalc = MSR/MSr. Order to be significant, Fcal > Ftab to 95% of the confidence limit.

independents variables studied in the process. The chart that The separation of terpenes via TLC revealed a total of seven
represents the yield of extracts Pitanga seeds, generated by the groups of different substances (C1–C7, Fig. 2). However, not all
quadratic model (Fig. 1b), the optimal region (in red) is that spots appear in all extracts, as the compound 7 which will not
where higher values of P were applied for the entire range appear in the extracts from the essay E1, E2, E3 and E5. The
of T studied. To operate the equipment with low tempera- qualitative analyses of terpenes by TLC revealed the presence
tures leads to obtaining a good yield and still maintain the of these compounds in all extracts tested, the retention factor
integrity of thermosensitive substances. Thus, the extraction (Rf) was measured for each compounds (Table 4). Retention
with supercritical CO2 can present a great advantage over con- factor is the ratio among the distance moved up by the com-
ventional methods, in which sometimes must, be used at high ponent and the distance moved by mobile phase (Gotkiewicz,
temperatures. Moreover, with use of kind temperatures, power 1981). Considering the gas chromatography, retention factor
consumption is less which is an economic consideration to be can be obtained from the relationship between the substance
considered. retention time and the total time of chromatographic run
(Collins et al., 1995). In addition to TLC analysis, extracts from
SFE was also analyzed by GC–MS whose retention times of
3.2. Refractive index (RI) and thin layer
the principal components are listed in Table 5. Specifically, for
chromatography (TLC)
the analysis of GC–MS, the relationship between the retention
time of each compound with total run time of 67 min resulted
The refractive index of a substance is the ratio between
in a Rf of 0.42 for the ␥-elemene and 0.50 to germacrone.
the speed of light in vacuum and the speed of light in the
Comparing the Rf from two chromatographic methods it
substance (Jorge and Lopes, 2003) and one of the most impor-
was observed that the ␥-elemene Rf from GC–MS is simi-
tant parameters of the optical material, in performance an
lar to Rf of the C4 Rf (4th band, Fig. 2 and Table 4) and the
important role in chemical and physical substances charac-
germacrone Rf approaches to C5 Rf. This relation could indi-
terizations. A precise knowledge of refractive index shows
cate that these components identified in gas chromatography
the purity and reproducibility of the material for optical
could be present in TLC plates. Germacrone and ␥-elemene
instruments (Samedov, 2006). Table 2 shows the refractive
are sesquiterpenes. This class of compounds identified for
index of the cherry seeds extracts at 20 ◦ C, which varied from
Wagner and Bladt (2009), using the same methodology, pre-
1.503 to 1.519. The refractive indices for the test E9, E10 and
sented similar Rfs those obtained for Pitanga seeds extracts
E11 which were obtained in the same operational conditions
from SFE.
presented the similarity which proves the reproducibility of
Terpenes are characteristic compounds in essential oils
measurements.
and it in combination with the refractive index values sug-
The refractive index is a physical property that may indi-
gest that the supercritical extracts from Brazilian cherry seeds
cate the composition of an extract. For essential oils, normally
are essential oil behavior. Essential oils are complex mixtures
range from 1.450 to 1.590 at 20 ◦ C (Baser and Buchbauer, 2010).
which may contain 100 or more organic compounds. Con-
Whereas seeds usually contain lipids in their composition
stituents can belong to several classes of compounds, but
and, the aim of this study was to obtain oil from Pitanga
terpenes and phenylpropenes are the most common classes of
seed rich in volatile compounds, the measurement of this
compounds in essential oils. The most frequently terpenes in
physical property has been chosen for this characterization.
essential oils are the mono, di and sesquiterpenes, diterpenes
So, the refractive index measured shows that the compo-
are the minor constituents (Castro et al., 2004). Presently, about
nents present in the supercritical extracts from Brazilian
3000 essential oils are known, of which 300 are commercially
cherry seeds are characteristic of essential oil for all extrac-
important, especially for the pharmaceutical, dental, food,
tion conditions. This indicates that no operational condition
agronomic, health, cosmetics and perfume industries for var-
studied favored the lipids extraction. Rosemary essential oil
ious purposes (Bakkali et al., 2008).
obtained by microwave and hydrodistillation presented RI
TLC analyses using the same mobile phase with DPPH
values of 1.468 and 1.470, respectively (Bousbia et al., 2009).
free radical as developer was made in order to determine
Thyme essential oil obtained from hydrodistillation assisted
spots antioxidant activity. In this analysis it was possible to
by microwave and conventional hydrodistillation presented
visualize yellow spots, an indicative of inhibition of the free
refractive index similar to extracts from Brazilian cherry seeds
radical (Fig. 2). Thus, it is possible that these terpenes have
of 1.503 at 20 ◦ C (Golmakani and Rezaei, 2008).
370 food and bioproducts processing 9 4 ( 2 0 1 5 ) 365–374

Table 4 – Retention factor of Pitanga seeds supercritical extracts from thin layer chromatography.
Extract C1 C2 C3 C4 C5 C6 C7

E1 0.21 0.29 0.41 0.55 0.68 – –

E2 0.21 0.32 0.42 0.55 0.65 – –
E3 0.21 0.33 0.42 0.55 0.65 – –
E4 0.21 0.32 0.43 0.55 0.65 – 0.81
E5 0.21 0.33 0.43 0.55 0.68 – –
E6 0.21 0.32 0.43 0.55 0.68 – 0.86
E7 0.21 0.32 0.41 0.56 0.71 – 0.88
E8 0.19 0.28 0.40 0.56 0.65 – 0.86
E9 0.19 0.28 0.40 0.56 0.65 – 0.86
E10 0.20 0.28 0.40 0.56 0.64 – 0.87
E11 0.20 0.28 0.40 0.56 0.65 – 0.86
SOXH 0.20 0.26 – – – – 0.90
SC-EtOH – – – – – 0.84 0.91
PURIFI 0.20 0.26 – – – 0.84 0.91

Table 5 – Volatile compounds of Pitanga (Eugenia uniflora) seeds supercritical extracts identified by GC–MS analysis.
Peak Compound RT (min) Peak area in the extracts (%) AI KI KI*

1 2 3 4 5 6 7 8 9 10 11

1 Isosericenin 26.23 4.27 4.62 5.11 4.50 3.32 5.13 4.36 4.34 4.25 4.64 3.47 1450 1454
2 ␥-Elemene 28.53 39.47 39.02 38.17 46.14 36.70 42.93 40.52 41.99 43.86 43.52 43.83 1479 1481 1436a
3 Impurity 29.43 – – – – – – – – – – –
4 N.I. 31.36 3.23 3.97 4.87 2.53 4.42 3.28 3.48 3.33 2.84 3.07 3.26 1617 1618
5 Furanodiene 33.60 15.01 15.20 14.84 13.19 12.33 15.62 13.62 13.46 13.27 14.63 11.67 1648 1650 1666b
6 Germacrone 33.72 38.02 37.19 37.01 33.64 43.22 33.04 38.02 36.87 35.78 34.13 37.77 1648 1652 1658c

RT, retention time; KI, Kovatz index determined using C9–C22 hydrocarbon as external reference standard; KI*, Kovatz index of literature.
a
Costa et al. (2010).
b
Chang et al. (2011).
c
Marongiu et al. (2005).

Fig. 2 – Chromatogram of Pitanga seed extracts in silica gel Plate 60 F254 , mobile phase toluene and ethyl acetate (93:7, v:v)
with revealers different. (Above) Anisaldehyde as spray reagent for terpenes detection. (Below) As spray reagent DPPH
solution for antioxidant activity detection. SOXH, Soxhlet extract; SC-EtOH, supercritical extract obtained with ethanol;
PURIFI, purified fraction of SC-EtOH extract.
 food and bioproducts processing 9 4 ( 2 0 1 5 ) 365–374
Fig. 3 – Total ion chromatogram of the volatile compounds present in Pitanga seed extracts. Detail in the expanded section,
where the main peaks are more evident.
antioxidant activity, as so as indicated to Souza et al. (2007),
which assumed the thermotolerance, photoprotective and
antioxidant actions of the monoterpenes from plants, espe-
cially related to its ability to capture oxygen radicals from the
photosynthetic process. In a work with has used essential oils
of turmeric, the analysis of TLC with DPPH radical (0.04%)
as developer indicated that the standards used (curzerene,
germacrone and furanodienone) showed antioxidant activ-
ity (Zhao et al., 2010a). The identification of the compounds
present in essential oil allows determining the contribution of
the antioxidant activity of each compound.
3.3. Volatile composition in the extracts
Fig. 3 shows the ion chromatogram (TIC) of the supercritical
extract from Brazilian cherry seeds obtained at the cen-
tral point of the CCD, also with an enlargement of the TIC
region where the main peaks are concentrated. In these chro-
matograms, the peaks that appear more prominent in the
retention times of 28.53 and 33.72 min, are the ␥-elemene and
germacrone, respectively. For each extract, Table 5 shows the
area percentage for each volatile found.
Among the supercritical extracts, the concentrations of
major volatile, ␥-elemene and germacrone, considering the
sum of the area percentage of the peaks in the calculation,
ranged from 75.18 to 81.60 g/100 g. For germacrone, the con-
centration varied with the different operational conditions
applied, from 33.04 g/100 g for the essay 6 to 38.02 g/100 g for
the essays 1 and 7. Specifically for the ␥-elemene, the variation
in the concentration was from 36.70 g/100 g for the essay 5 to
46.14 g/100 g for the essay 4.
The Kovatz index (KI) of each compound was calculated as
a tool for the identification in addition to the mass spectra
library comparison. This index was calculated and compared
with the values reported in the literature, as shown in Table 5.
The KI values calculated for the volatile compounds from
Brazilian cherry seeds are similar to the references.
The furanodiene, third compound more intense in the
extracts was indicated as curzerene by the library software,
with 71–77% of similarity. The KI calculated showed a large
difference with the KI values of the curzerene available in the
literature, which was around 1470 and the value found in this
study was 1650. A recent study examined the influence of the
conditions of GC analysis on the quantification of thermosen-
sitive volatile compounds in the essential oil from Brazilian
cherry leaves (Chang et al., 2011). It was observed a structural
conversion of furanodiene in curzerene under the conditions
371
of high temperature and column similar to this study. When
slight temperature conditions were used by these authors this
conversion was not observed. According to their work, it is
possible to have partial concentrations of each compound
(furanodiene and curzerene) in the extracts. However, taking
account the KI calculated and compared with those reported
in the literature, it was considered that the substance in the
volatile fraction from the Brazilian cherry seeds extracts is the
furanodiene.
Germacrone, furanodiene and ␥-elemene are germacranes
sesquiterpenes, which are assigned various effects, such
as intermediates of biosynthesis of several sesquiterpenes
such as pheromones, anti-inflammatory, anti-cancer agents,
antiplasmodial activity, antibacterial and antifungal, phyto-
toxic activity, inhibiting tumor necrosis, nematicidal activity,
cytotoxic activity, tyrosinase inhibitory activity and other
diverse biological activities (Adio, 2009; Zhong et al., 2011).
Essential oil from Curcuma zedoaria is administered as a
medicament and the germacrone, appearing in its composi-
tion as a major compound, and for this it is a key critical
component that is measured in the quality control of this
essential oil (Zhao et al., 2010b). Furanodiene has been iden-
tified as having analgesic, anti-inflammatory and protective
effects on liver against tumor. The furanodiene not only
has direct antitumor action, but may also improving the
immune function of the body. This compound shows the
potential to be a clinically effective antitumor agent due
to its potent anti-cancer effect and low toxicity (Ba et al.,
2009).
Response surface analysis (RSA) was applied as a tool to
understand how the operational variables (P and T) influenced
the presence of a compound in the extract. It is important
to arrive at operating conditions that support the fraction-
ation of these compounds. The concentration (mg/kg) of
the two main components in the extract (germacrone and
␥-elemene) presented as limonene equivalent (LimEq) was
determined for the 11 essays (Table 2). For the germacrone,
the Pareto diagram (Fig. 4a) shows that the interaction of P
and T and also P showed a significant influence to obtain
extracts rich in this sesquiterpene. For the ␥-elemene, the
Pareto diagram shows that only the interaction between pres-
sure and temperature presented a significant effect in the
model (Fig. 4b).
Table 6 shows the mathematical models, linear and
quadratic, which relate the behavior of the ␥-elemene and
germacrone concentration present in extracts as a function
of P and T. Although no models are able to adequately
372 food and bioproducts processing 9 4 ( 2 0 1 5 ) 365–374

Table 6 – Analysis of variance for first and second order factorial experimental design for Brazilian cherry seeds extract
germacrone and ␥-elemeno (mg/kg LimEq).
Source of variation Sum square Degree of freedom Mean square Fcalc Ftab
(SS) (DF) (MS)

Germacrone
Linear model (first order) – R = 18.673 − 5.615P–7.960PT
Model (R) 509.452 2 180.241 98.76 6.94
Residue (r) 5.182 4 1.825
Total 514.634 6
Quadratic model (second order) – R = 14.363 − 2.660P + 3.475P2 − 7.960PT
Model (R) 534.547 3 178.183 240.78 4.35
Residue (r) 5.182 7 0.740
Total 539.729 10

␥-Elemene
Linear model (first order) – R = 21.457 − 7.185PT
Model (R) 410.917 1 410.917 140.87 6.61
Residue (r) 14.584 5 2.9168
Total 425.501 6
Quadratic model (second order) – R = 18.072 − 7.185PT
Model (R) 505.239 1 505.239 311.68 5.12
Residue (r) 14.585 9 1.621
Total 519.824 10

Fcalc = MSR/MSr. Order to be significant, Fcal > Ftab to 95% of the confidence limit.

predict the experimental values for the ␥-elemene, analysis of

Fig. 4 – Pareto chart with the effects of independent
variables (P and T) in first order model, on the
concentration of (a) germacrone and (b) ␥-elemene in
Pitanga seed extracts.
variance for linear and quadratic model (ANOVA) (Table 6), in
accordance with the Pareto diagram, show, by comparing the
calculated and tabulated values for F (Fcalc and Ftab ) with 5%
probability, that at least one of the variables influencing the
concentration of sesquiterpenes in the supercritical extract
from Brazilian cherry seeds, since Fcalc was considerably
larger than the Ftab for these two compounds.
In the evaluation of the extraction independent variables
(P and T) by the second order model for the concentration
of these two sesquiterpenes, the response surface analysis
indicated, in the range studied, that high temperatures and
low pressures provide a higher concentration of these volatile
in the extracts (Fig. 5). These conditions are reversed to the
extracts yields. High temperature favors the volatiles extrac-
tion, indicating the influence that in these conditions, the
influence of the vapor pressure of the compounds may be
the most important factor to be considered in supercriti-
cal fluid extraction (Table 3). Although it could not possible
promptly to evaluate the influence of high temperature and
low densities, (since the temperature range studied was rel-
atively low, from 30 to 60 ◦ C approximately, it was observed
that the experiment whose lower CO2 densities resulted
in the highest concentration of the two main sesquiter-
penes at 55 ◦ C. This temperature, in the CCD applied in
this study is high, but compared to other extraction meth-
ods such as steam distillation, it could still be considered
mild.
In fact, monoterpene hydrocarbons such as non-polar com-
pounds of relatively low molecular weight are very soluble
in SC-CO2 and therefore it is highly possible that they may
be completely extracted from the matrix. Due to its high
volatility, they are partially separated from the gaseous CO2
and precipitate on the collector and a second part escapes
dissolved in the gas. Thus, especially for the monoterpenes
is more correct to measure the collection efficiency of the
extract than the extraction yields (Bocevska and Sovová,
2007).
 food and bioproducts processing 9 4 ( 2 0 1 5 ) 365–374
Fig. 5 – Response surface plots by the second order model
that shows the effects of independent variables (P and T) in
the concentration of (a) germacrone and (b) ␥-elemene in
the Pitanga seed extracts.
4. Conclusion
The yield of Pitanga seeds extracts was higher when car-
ried out under high pressures conditions and the temperature
not influenced this response. The volatile profile performed
by GC–MS indicated as major components in the supercrit-
ical extracts from Brazilian cherry seeds the sesquiterpenes
germacrone and ␥-elemene, which play numerous biological
activities such as literature reports. The presence of volatile
terpenes was previously demonstrated by TLC, which also
showed indication of antioxidant activity. In this research
group, more investigation follow in order to study the bioac-
tivity of these extracts, presented in this work motivated the
studies of the extracts activities in progress.
373
Acknowledgements
The authors would like to thank FAPESP (State of São Paulo
Research Foundation, Brazil) for financial support under the
research projects 2008/00148-0 and CAPES for the scholarship
the master student.
References
Adio, A.M., 2009. Germacrenes A–E and related compounds:
thermal, photochemical and acid induced transannular
cyclizations. Tetrahedron 65, 1533–1552.
Adams, R.P., 2009. Identification of Essential Oil Components by
Gas Chromatography/Mass Spectrometry, 4th ed. Allured
books, Carol Stream.
Amorim, A.C.L., Lima, C.K.F., Hovella, A.M.C., Miranda, A.L.P.,
Rezende, C.M., 2009. Antinociceptive and hypothermic
evaluation of the leaf essential oil and isolated terpenoids
from Eugenia uniflora L. (Brazilian Pitanga). Phytomedicine 16,
923–928.
Andreo, D., Jorge, N., 2006. Antioxidantes naturais: técnicas de
extração. Bol. CEPPA 24 (2), 319–336.
Association of Analytical Communities, 1995. Official Methods of
Analysis of AOAC International. AOAC International,
Arlington.
American Society of Agricultural and Biological Engineers, 2008.
Method of Determining and Expressing Fineness of Feed.
Materials by Sieving. S319.4, MI.
Auricchio, M.T., Bugno, A., Barros, S.B.M., Bacchi, E.M., 2007.
Atividades antimicrobiana e antioxidante e toxicidade de
Eugenia uniflora. Lat. Am. J. Pharm. 26 (1), 76–81.
Ba, Z.Z., Zhen, Y., Zhang, H., Sun, X., Lin, D., 2009. Potential
anti-cancer activity of furanodiene. Chin. J. Cancer Res. 21 (2),
154–158.
Bagetti, M., Facco, E.M.P., Rodrigues, D.B., Vizzotto, M., Emanuelli,
T., 2009. Antioxidant capacity and composition of Pitanga
seeds. Ciênc. Rural 39 (8), 2504–2510.
Bakkali, F., Averbeck, S., Averbeck, D., Idaomar, M., 2008.
Biological effects of essential oils – a review. Food Chem.
Toxicol. 46, 446–475.
Baser, K.H.C., Buchbauer, G. (Eds.), 2010. Handbook of Essential
Oils: Science, Technology, and Applications. CRC Press, Boca
Raton.
Berna, A., Tárrega, A., Blasco, M., Subirats, S., 2000. Supercritical
CO2 extraction of essential oil from orange peel; effect of the
height of the bed. J. Supercrit. Fluids 18, 227–237.
Bezerra, J.E.F., Lederman, I.E., Silva-Júnior, J.F., Alves, M.A., 2004.
Comportamento da pitangueira (Eugenia uniflora L.) sob
irrigação na região do vale do rio Moxotó, Pernambuco. Rev.
Bras. Frutic. 26 (1), 177–179.
Bocevska, M., Sovová, H., 2007. Supercritical CO2 extraction of
essential oil from yarrow. J. Supercrit. Fluids 40, 360–367.
Bousbia, N., Vian, M.A., Ferhat, M.A., Petitcolas, E., Meklati, B.Y.,
Chemat, F., 2009. Comparison of two isolation methods for
essential oil from rosemary leaves: hydrodistillation and
microwave hydrodiffusion and gravity. Food Chem. 114,
355–362.
Castro, H.G., Oliveira, L.O., Barbosa, L.C.A., Ferreira, F.A., Silva,
D.J.H., Mosquim, P.R., Nascimento, E.A., 2004. Teor e
composição do óleo essencial de cinco acessos de menstrato.
Quím. Nova 27 (1), 55–57.
Chang, R., Morais, S.A.L., Napolitano, D.R., Duarte, K.C., Guzman,
V.B., Nascimento, E.A., 2011. A new approach for quantifying
furanodiene and curzerene. A case study on the essential oils
of Eugenia uniflora (Pitangueira) leaves. Rev. Bras. Farmacogn.
21 (3), 392–396.
Collins, C.H., Braga, G.L., Bonato, P.S., 1995. Introdução a métodos
cromatográficos, 6th ed. Editora da Unicamp, Campinas.
Costa, D.P., Alves Filho, E.G., Silva, L.M.A., Santos, S.C., Passos,
X.S., Silva, M.R.R., Seraphin, J.C., Ferri, P.H., 2010. Influence of
fruit biotypes on the chemical composition and antifungal
374 food and bioproducts processing 9 4 ( 2 0 1 5 ) 365–374
activity of the essential oils of Eugenia uniflora leaves. J. Braz.
Chem. Soc. 21 (5), 851–858.
Danh, L.T., Mammucari, R., Truong, P., Foster, N., 2009. Response
surface method applied to supercritical carbon dioxide
extraction of Vetiveria zizanioides essential oil. Chem. Eng. J.
155, 617–626.
Empresa Brasileira de Pesquisa Agropecuária, Antunes, L.E.C.,
Raseira, M.C.B., 2006. III Simpósio nacional do morango. II
Encontro sobre pequenas frutas e Frutas nativas do Mercosul.
Documentos 171. Pelotas.
Golmakani, M.T., Rezaei, K., 2008. Supercritical fluid extraction in
plant essential and volatile oil analysis. Food Chem. 109,
925–930.
Gotkiewicz, W., 1981. TLC as a pilot technique for the
optimization of gradient HPLC I. Experimental verification of a
graphical method for the optimization of stepwise gradient
elution. Chromatographia 14 (7), 411–414.
Huang, F.H., Li, M.H., Lee, L.L., Starling, K.E., Chung, F.T.H., 1985.
An accurate equation of state for carbon dioxide. J. Chem. Eng.
Jpn. 18 (6), 490–496.
Instituto Adolfo Lutz, 2008. Métodos físico-químicos para análise
de alimentos, 4th ed, São Paulo.
Jorge, N., Lopes, M.R.V., 2003. Avaliação de óleos e gorduras de
frituras coletados no comércio de São José do Rio Preto-SP.
Aliment. Nutr. 14 (2), 149–156.
Laroze, L.E., Díaz-Reinoso, B., Moure, A., Zúñiga, M.A.,
Domínguez, H., 2010. Extraction of antioxidants from several
berries pressing wastes using conventional and supercritical
solvents. Eur. Food Res. Technol. 231, 669–677.
Lhuillier, A., Fabre, N., Moyano, F., Martins, N., Claparols, C.,
Fourasté, I., Moulis, C., 2007. Comparison of flavonoid profiles
of Agauria salicifolia (Ericaceae) by liquid chromatography-UV
diode array detection–electrospray ionisation mass
spectrometry. J. Chromatogr. A 1160, 13–20.
Lopes Filho, G., Rosso, V.V., Meireles, M.A.A., Rosa, P.T.V., Oliveira,
A.L., Mercadante, A.Z., Cabral, F.A., 2008. Supercritical CO2
extraction of carotenoids from Pitanga fruits (Eugenia uniflora
L.). J. Supercrit. Fluids 46, 33–39.
Luzia, D.M.M., Bertanha, B.J., Jorge, N., 2010. Sementes de pitanga
(Eugenia uniflora L.): potencial antioxidante e perfil de ácidos
graxos. Rev. Inst. Adolfo Lutz 69 (2), 175–180.
Malaman, F.S., Moraes, L.A.B., West, C., Ferreira, N.J., Oliveira,
A.L., 2011. Supercritical fluid extracts from the Brazilian
cherry (Eugenia uniflora L.): relationship between the extracted
compounds and the characteristic flavour intensity of the
fruit. Food Chem. 124, 85–92.
Marongiu, B., Piras, A., Porcedda, S., Scorciapino, A., 2005.
Chemical composition of the essential oil and supercritical
CO2 extract of Commiphora myrrha (Nees) Engl. and of Acorus
calamus L. J. Agric. Food. Chem. 53, 7939–7943.
Marsal, A., Celma, P.J., Cot, J., Cequier, M., 2000. Supercritical CO2
extraction as a clean degreasing process in the leather
industry. J. Supercrit. Fluids 16, 217–223.
Marsili, R. (Ed.), 2002. Flavor, Fragrance, and Odor Analysis.
Marcel Decker, New York.
Mélo, E.A., Lima, V.L.A.G., Nascimento, P.P., 1999. Formulação e
avaliação físico-química e sensorial de geléia mista de
pitanga (Eugenia uniflora L.) e acerola (Malpighia sp). Bol. CEPPA
17 (1), 33–44.
Ogunwande, I.A., Olawore, N.O., Ekundayo, O., Walker, T.M.,
Schmidt, J.M., Setzer, W.N., 2005. Studies on the essential oils
composition, antibacterial and cytotoxicity of Eugenia uniflora
L. Int. J. Aromather. 15, 147–152.
Oliveira, A.L., Destandau, E., Fougère, L., Lafosse, M., 2014.
Isolation by pressurized fluid extraction (PFE) and
identification using CPC and HPLC/ESI/MS of phenolic
compounds from Brazilian cherry seeds (Eugenia uniflora L.).
Food Chem. 145, 522–529.
Oliveira, A.L., Lopes, R.B., Cabral, F.A., Eberlin, M.N., 2006. Volatile
compounds from Pitanga fruit (Eugenia uniflora L.). Food Chem.
99, 1–5.
Oliveira, A.L., Kamimura, E.S., Rabi, J.A., 2009. Response surface
analysis of extract yield and flavour intensity of Brazilian
cherry (Eugenia uniflora L.) obtained by supercritical carbon
dioxide extraction. Innovative Food Sci. Emerging Technol. 10,
189–194.
Oliveira, D.A., (Dissertação de Mestrado) 2010. Caracterização
fitoquímica e biológica de extratos obtidos de bagaço de uva
(Vitis vinifera) das variedades merlot e syrah. Centro
Tecnológico da Universidade Federal de Santa Catarina,
Florianópolis, SC, Brasil, 211 pp.
Pino, J.A., Bello, A., Urquiola, A., Aguero, J., Marbot, R., 2003. Fruits
volatiles of Cayena Cherry (Eugenia uniflora L.) from Cuba. J.
Essent. Oil Res. 15, 70–71.
Salgin, U., 2007. Extraction of jojoba seed oil using supercritical
CO2 + ethanol mixture in green and high-tech separation
process. J. Supercrit. Fluids 39, 330–337.
Samedov, F., 2006. Laser-based optical facility for determination
of refractive index of liquids. Opt. Laser Technol. 38, 28–36.
Serra, A.T., Seabra, I.J., Braga, M.E.M., Bronze, M.R., Sousa, H.C.,
Duarte, C.M.M., 2010. Processing cherries (Prunus avium) using
supercritical fluid technology. Part 1: Recovery of extract
fractions rich in bioactive compounds. J. Supercrit. Fluids 55,
84–191.
Silva, S.M., 2006. Pitanga. Rev. Bras. Frutic. 28 (1), 1.
Souza, T.J.T., Apel, M.A., Bordignon, S., Matzenbacher, N.I.,
Zuanazzi, A.S., Henriques, A.T., 2007. Composição química e
atividade antioxidante do óleo volátil de Eupatorium
polystachyum DC. Rev. Bras. Farmacogn. 17 (3), 368–372.
Wagner, H., Bladt, S., 2009. Plant Drug Analysis – A Thin Layer
Chromatography Atlas. Springer, London.
Zhang, S., Zu, Y., Fu, Y., Luo, M., Liu, W., Li, J., Efferth, T., 2010.
Supercritical carbon dioxide extraction of seed oil from yellow
horn (Xanthoceras sorbifolia Bunge.) and its anti-oxidant
activity. Bioresour. Technol. 101, 2537–2544.
Zhao, J., Zhang, J., Yang, B., Lv, G., Li, S., 2010a. Free radical
scavenging activity and characterization of sesquiterpenoids
in four species of Curcuma using a TLC bioautography assay
and GC–MS analysis. Molecules 15, 7547–7557.
Zhao, Y., Wang, C., Chow, A.H.L., Ren, K., Gong, T., Zhang, Z.,
Zheng, Y., 2010b. Self-nanoemulsifying drug delivery system
(SNEDDS) for oral delivery of Zedoary essential oil:
formulation and bioavailability studies. Int. J. Pharm.
(Amsterdam, Neth.) 383, 170–177.
Zhong, Z., Chen, X., Tan, W., Xu, Z., Zhou, K., Wu, T., 2011.
Germacrone inhibits the proliferation of breast cancer cell
lines by inducing cell cycle arrest and promoting apoptosis.
Eur. J. Pharmacol. 667, 50–55.