J.

of Supercritical Fluids 72 (2012) 270–277

Contents lists available at SciVerse ScienceDirect

The Journal of Supercritical Fluids

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

Supercritical fluid extraction of corn germ oil: Study of the influence of process
parameters on the extraction yield and oil quality
Sara Rebolleda, Nuria Rubio, Sagrario Beltrán, M. Teresa Sanz ∗ , María Luisa González-Sanjosé
Department of Biotechnology and Food Science, University of Burgos, Plaza Misael Bañuelos s/n, 09001 Burgos, Spain

a r t i c l e i n f o a b s t r a c t

Article history: The supercritical fluid extraction of corn germ oil has been studied in this work. Extractions were carried
Received 11 July 2012 out at different pressure, temperature and flow rate to analyze the influence of these variables on the
Received in revised form 3 October 2012 extraction kinetics and the oil quality obtained. Extraction curves are initially linear with a slope close to
Accepted 4 October 2012
the oil solubility value in supercritical CO2 . Based on these results a mathematical model was successfully
applied to describe the extraction curves. Characterization of supercritical crude corn oil was performed
Keywords:
by determining some physical parameters such as refraction index, density and color. Additionally, the
Supercritical fluid extraction
fatty acid composition, neutral lipids, the content of tocopherols, acid index, peroxide value, antioxidant
Corn germ oil
Tocopherol
capacity and the oxidative stability were determined in the corn oil extracted. Fatty acid composition was
Sovová’s model compared with that for crude germ oil and no significant differences between the oils extracted by both
methods were found. Oxidative stability test using the Rancimat showed that supercritical CO2 extracted
corn oil is less protected against oxidation than n-hexane extracted oils.
© 2012 Elsevier B.V. All rights reserved.

1. Introduction polar triglycerides could be separated from high polar compounds

such as water and free fatty acids.
Traditionally, the extraction of corn germ oil has been done by Vigh et al. [6] reported extraction curves of wet corn germ oil
physical and chemical methods. In these conventional methods, the with SC-CO2 in an interval of pressure (27.0–33.0 MPa) and tem-
oil is removed from the milled germ using a conditioning (heat- perature (42–78 ◦ C) and at particle size <0.8 mm or 0.8–1.4 mm.
ing) process, followed by mechanical expelling (prepress) and in Based on a second order orthogonal design, Vigh et al. [6] concluded
some cases ending up with hexane extraction. Extrusion has been that for the smaller particle size range, application of low pressures
also employed to prepare the germ for solvent extraction leading and high temperatures should be avoided. Rónyai et al. [7] studied
a crude corn oil of high quality and high yield [1]. After oil extrac- the SC-CO2 extraction of wet-milled corn germ oil with cosolvents
tion, the bagasse obtained as raffinate is normally used for animal (ethyl alcohol: from 0% to 10% by weight in CO2 ) at a constant
nutrition. pressure and temperature (30 MPa and 42 ◦ C). The extraction time
The supercritical fluid extraction (SFE) of corn gem oil has been was reduced due to the higher solubility of corn oil with increasing
studied by several authors. List et al. [2,3] and Christianson et al. ethanol concentration although the amount of phospholipids in the
[4] compared the quality of crude oils obtained from dry and wet oil was higher.
milled corn germ by SFE at 50–90 ◦ C and 55–83 MPa and by con- In addition to oil, SFE of corn germ also results in an interesting
ventional extraction methods. The oil obtained by supercritical defatted corn germ flour, due to its low fat content and lower perox-
CO2 extraction exhibited lower refining loss and lighter color. List idase activity when compared to hexane extracted corn germ flour,
et al. [2] found that, in general, crude oil quality was unaffected which lead to a larger storage stability and better flavor [4,8,9].
by extraction conditions. The levels of tocopherols present in SC- Rónyai et al. [7] found that the emulsifying, foaming and absorp-
CO2 extracted corn oil are similar to those obtained by conventional tion properties of the defatted meals and its protein isolates were
methods [2,3]. However the phospholipids are almost absent in the better when using ethyl alcohol as cosolvent.
SC-CO2 oils which is an advantage from a processing point of view. The aim of this work is the study of the influence of some extrac-
Wilp and Eggers [5] proposed an extraction process followed by tion parameters such as pressure, temperature and solvent flow
a fractionation separation step to improve the oil quality since low rate on the corn germ oil extraction rate. New extraction curves
have been obtained than those reported in the literature. In addi-
tion, the quality and stability of corn germ oil obtained by SFE, with
∗ Corresponding author. Tel.: +34 947 258810; fax: +34 947 258831. and without fractionation, was evaluated under different extrac-
E-mail address: tersanz@ubu.es (M.T. Sanz). tion conditions. Quality and stability parameters not evaluated

0896-8446/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.supflu.2012.10.001
 S. Rebolleda et al. / J. of Supercritical Fluids 72 (2012) 270–277 271

Table 1
Nomenclature Experimental conditions in the SFE of corn germ oil with SC-CO2 .

Run p T Solvent flow rate Fractionation

as specific area between the regions of intact and bro- (MPa) (◦ C) (kg CO2 /h)
ken cells (m−1 )
R1 45.0 ± 2.0 40 ± 1.5 8.5 ± 0.8 No
C1 , C2 fitting parameters R2 45.0 ± 2.0 63 ± 1.5 8.8 ± 0.8 No
e extraction yield (kg extract/kg insoluble solid) R3 45.5 ± 2.0 79 ± 2 9.5 ± 0.9 No
E extract (kg) R4 45.0 ± 1.8 85 ± 2.5 8.0 ± 0.9 No
ks solid-phase mass transfer coefficient (s−1 ) R5 30.0 ± 1.5 80 ± 1.5 10.0 ± 0.7 No
R6 21.0 ± 0.5 39 ± 1 9.0 ± 0.7 No
n number of experimental data
R7 52.5 ± 2.0 80 ± 2 6.0 ± 0.5 No
Nm charge of insoluble solid (kg) R8 44.2 ± 2.0 78 ± 2 3.9 ± 0.5 No
O.F. objective function R9 45.9 ± 2.0 56 ± 2 10.0 ± 0.9 No
Q solvent flow rate (kg h−1 ) R10 45.5 ± 1.7 57 ± 1.5 8.0 ± 0.7 No
R11 44.0 ± 1.5 84 ± 1.5 6.0 ± 0.7 No
q relative amount of the passed solvent (kg solvent/kg
R12 43.1 ± 2.0 84 ± 2 5.0 ± 0.5 No
insoluble solid) R13 50.0 ± 1.5 35 ± 1.5 11.0 ± 0.7 Yes
qc relative amount of the passed solvent when all the R14 48.0 ± 1.5 86 ± 1.5 7.0 ± 0.9 Yes
solute in broken cells has been extracted (kg sol- R15 25.8 ± 2.0 85 ± 1.5 9.0 ± 0.9 Yes
vent/kg insoluble solid)
r grinding efficiency (fraction of broken cells)
t extraction time (h) a subsequent fractionation in two separators installed in series
xu concentration in the untreated solid (kg solute/kg were carried out. The first separator was maintained at 10.0 MPa
insoluble solid) and 40 ◦ C in order to recover the less soluble compounds (triacyl-
ys solubility (kg solute/kg solvent) glycerides), and the second one at 4.0 MPa and a temperature of
40 ◦ C. Most of the water was found to be recovered in the second
separator.
before in corn germ oil extracted with SC-CO2 , such as antioxidant
capacity, have been determined. 2.3. Analytical methods

2. Experimental 2.3.1. Physical oil properties

Some important parameters like refraction index (Milton Roy
2.1. Raw material abbe-type refractometer), density (densimeter model DMA 5000,
Anton Paar) and color (CIELab parameters) were evaluated in
The raw material used in this work was corn germ with an the different oils extracted. CIELab parameters were calculated
average moisture content of 6.0 ± 0.4% and 46 ± 3% average fat con- automatically by a suitable programme installed in a Beckman
tent as determined by Soxhlet extraction with petroleum ether. DU-650 spectrophotometer with diode-array of UV–vis (Analytical
Regarding the moisture content, Christianson et al. [4] reported Development Center, Beckman Instruments Inc., 1995), using the
that, at this level, it plays a minor role in SC-CO2 extraction of corn illuminant D65 (daylight source) and a 10◦ standard observer (per-
germ oil not affecting the extraction efficiency or the ultimate oil ception of a human observer) following the CIE recommendations
yield. The corn germ was milled in a coffee grinder to a particle size [11].
ranging from 0.5 mm to 1 mm.
2.3.2. Determination and quantification of fatty acids profile
2.2. Supercritical fluid extraction equipment and procedure The fatty acids profile was determined by the AOAC method [12].
The fatty acid methyl esters were firstly prepared and then analyzed
The extraction experiments were carried out in a semi-pilot by gas chromatography (GC) in a Hewlett Packard gas chromato-
SFE-plant whose P&I diagram has been presented elsewhere [10]. graph (6890N Network GC System) equipped with an auto-sampler
The usual elements of an SFE-plant with solvent recycling were (7683B series) and a flame ionization detector (FID). The separation
installed, i.e.: pump, extractor, separator, heating and cooling sys- was carried out with helium (1.8 mL/min) as carrier gas. A fused sil-
tems and pressure dampers; rupture disks and safety valves were ica capillary column (OmegawaxTM-320, 30 m × 0.32 mm i.d.) was
installed for safety and instruments for measurement and control used. The column temperature was programmed starting at a con-
of the process parameters. The maximum specifications of the SFE- stant temperature of 180 ◦ C for 20 min, heated to 200 ◦ C at 1 ◦ C/min,
plant are: T = 200 ◦ C, p = 65.0 MPa and solvent flow, Q = 20 kg/h. held at 200 ◦ C for 1 min, heated again to 220 ◦ C at 5 ◦ C/min and
In a SFE experience, 350 g of corn germ were placed in the extrac- finally held at 220 ◦ C for 20 min. A split injector (50:1) at 250 ◦ C
tor mixed with an inert filling in order to avoid bed compactation. was used. The FID was also heated to 250 ◦ C. Most of the fatty
The extractor was later pressurized with CO2 (Carburos metálicos, acid methyl esters were identified by comparison of their retention
liquid CO2 ≥ 99.9%) up to the extraction pressure. Then, the solvent times with those of chromatographic standards (Sigma Chemical
was circulated at the desired extraction pressure and temperature, Co.). Their quantification was made by relating the peaks area to
with a certain solvent flow, F, and during a specific time, t. The sol- the area of an internal standard (methyl tricosanoate) as indicated
vent was continuously recycled to the extractor after removing the by the AOAC method [12]. Calibration curves were made for several
solute in the separator. pairs formed by the internal standard + several representative chro-
A total of fifteen experiments under the different extraction matographic standards in order to find the corresponding response
conditions reported in Table 1 were carried out. Runs 1–8 were per- factors.
formed to study the influence of temperature, pressure and solvent
flow rate on the extraction kinetics. Runs 9–12 were performed to 2.3.3. Determination of neutral lipids
evaluate the influence of the extraction temperature on oil quality, The total amount of neutral lipids was determined by liquid
runs 9 and 10 were carried out at 56 ◦ C, and runs 11 and 12 were car- chromatography [13] in a HPLC system (Agilent 1200) formed
ried out at 84 ◦ C. Additionally, three experiments (R13–R15) with by a quaternary pump and an auto-injector. Separations were
272 S. Rebolleda et al. / J. of Supercritical Fluids 72 (2012) 270–277

carried out at room temperature in a column (Lichrospher Diol 3. Results and discussion
5 mm, 4 mm × 250 mm) and detection was performed in an evapo-
rative light scattering detector (Agilent 1200 series) at 35 ◦ C and 3.1. Influence of process parameters on the extraction yield
0.35 MPa. The mobile phase consisted of (A) isooctane and (B)
methyl tert-butyl ether:acetic acid (99.9:0.1). The solvent gradi- Supercritical fluid extraction (SFE) of a solute from a solid raw
ent used was as follows: first, solvent A was flowing for 1 min, after material may involve three different stages: internal mass transfer,
that, solvent B was added in three steps, up to 10% in 10 min, to phase equilibrium and external mass transfer. Thus, oil extraction
44% in 22 min and to 100% in 30 min. Finally, the stationary phase yield may be highly affected by operational parameters such as
was rinsed with solvent A for 5 min. Total solvent flow rate was solid pretreatment, extraction pressure and temperature and sol-
kept constant at 1 mL/min all along the analysis. Calibration was vent flow rate.
carried out using standards of palmityl palmitate (99%), tripalmitin The effect of extraction temperature on the extraction yield was
(>99%), dipalmitin (99%), monopalmitin (99%), palmitic acid (99%) evaluated from 40 ◦ C to 85 ◦ C at a constant pressure of 45.0 MPa
and ␤-sitosterol in isooctane. (runs 1–4) and a SC-CO2 flow rate around 9 kg/h. The results are
shown in Fig. 1 where it can be observed that the higher the tem-
perature the higher the extraction rate, which may indicate that,
2.3.4. Determination and quantification of tocopherol profile at this pressure, the increase of oil vapor-pressure with tempera-
Tocopherols determination in corn germ oil was done using ture is more important than the decrease in SC-CO2 density. As it is
HPLC–DAD after isolation by solid phase extraction (SPE). well established in literature [16], an increase of seed oil solubility
Solid phase extraction. The silica cartridge (1000 mg/6 mL, Sep- with extraction temperature can be significant when the process
Pak® , Waters, Spain) was conditioned with 5 mL of n-hexane before is performed at pressures higher than 40 MPa, pressure at which a
the application of 1 mL of n-hexane oil solution (0.1 g/mL). The crossover behavior is usually observed in vegetable oils [16].
elution of the analytes was done with 5 mL of n-hexane, fol- Regarding the effect of pressure, several extraction curves were
lowed by 5 mL of n-hexane–diethylether (99:1, v/v) and 50 mL obtained for pressures from 21 MPa to 53 MPa at two different tem-
of n-hexane–diethylether (99:2, v/v). The collected fractions were peratures around 40 ◦ C and 80 ◦ C (runs 1, 6 and 3, 5 and 7) and a
evaporated under reduced pressure at 45 ◦ C. The dry residue SC-CO2 flow rate around 9 kg/h. The results are shown in Fig. 2a
obtained was dissolved in 1.5 mL of n-hexane for HPLC analysis. and b, respectively. In both cases, the extraction curves indicate
High performance liquid chromatography. Tocopherols were per- that, at a constant temperature, the higher the pressure the higher
formed according to a modification of the IUPAC method [14] the extraction rate, what may be attributed to the higher density
using an Agilent HPLC (series 1100) equipped with ChemSta- of SC-CO2 which leads to higher solvent power. An increase of oil
tion software, a degasser (G1322A), a quaternary pump (G1311A), solubility when extraction pressure is increased has been reported
an autosampler (G1329A), a column oven (G1316A) and a diodo for dry-milled corn germ [4] and other seed oils [8]. As it will be
array detector (G1315A). The column used was ACE 5 silica explained in Section 3.2, the first part of the extraction is controlled
250 mm × 4.6 mm. The mobile phase was 99% hexane (A):1% 2- by this thermodynamic parameter and it can be fitted to a straight
propanol (B). An isocratic gradient was used and the total run time line. Fig. 2b indicates that the extraction at the lowest pressure (run
was 15 min. The injection volume was 50 ␮l. All tocopherols were 6, 21 MPa) is very slow which may be due to the low solubility of
monitored at 296 nm at a flow rate of 1 mL/min. seed oil in SC-CO2 .
Individual compounds of ␣-, ␤-, ␥- and ␦-tocopherols were iden- The effect of solvent flow rate has been studied at a constant
tified and quantified using a calibration curve of the corresponding pressure and temperature of 45 MPa and 79 ◦ C, respectively. Two
standard compound. different solvent flow rates were applied, 4 and 9 kg CO2 /h (runs
8 and 3). From Fig. 3 it can be concluded that extraction curves
expressed in dependence on the solvent-to-feed ratio are not sig-
nificantly affected by SC-CO2 flow rate. This behavior supports the
2.3.5. Acid index (AI) and peroxide value (PV)
fact that solubility, but not external mass transfer, controls the
The acid index and the peroxide value of the oils have
extraction process.
been determinated according to Ca 5a-40 and Cd 8-53 AOCS
Methods, respectively. An automatic titrator Methrom 905 Titrando
was used.

2.3.6. Determination of antioxidant capacity: DPPH assay

Free radical scavenging capacity of corn germ oil was eval-
uated using 2,2-diphenyl-1-picrylhydrazyl radical (DPPH• ) [15].
Briefly, 4 mL of DPPH• solution (0.1 mM) was mixed with 1 mL of
an isooctane corn germ oil solution (0.02 g/mL). The absorbance
at 517 nm was measured (Hitachi U-2000 spectrophotometer)
against a blank of pure isooctane after the reaction was carried
out at ambient temperature and darkness for 30 min. Results were
expressed in mmol of BHT/g oil, using the relevant calibration curve
described.

2.4. Oxidative stability

The oxidative stability was evaluated using the rancimat test

Fig. 1. Influence of extraction temperature on corn germ oil extraction yield at a
that was performed with a Metrohm 743 Rancimat using a 1.5 g oil constant pressure of 45.0–46.0 MPa ( 40 ◦ C;  63 ◦ C; ♦ 79 ◦ C;  85 ◦ C). The solid
sample, a temperature of 110 ◦ C and an air flow rate of 9 L/h. lines correspond to the model of Sovová [17].
 S. Rebolleda et al. / J. of Supercritical Fluids 72 (2012) 270–277 273

components are generally involved in the extraction of a seed oil

[18]. In the model of Sovová the extraction yield is expressed as:
E
e= (1)
Nm
where E is the amount of extract (kg) and Nm the charge of insoluble
solid (kg) in the extractor. The dimensionless amount of solvent
consumed is obtained by:
Qt
q= (2)
Nm
where Q is the solvent flow rate (kg/h) and t the extraction time
(h). Based on this model, the extraction curves consist of two parts.
During the first one, the easily accessible solute from broken cells
is transferred directly to the fluid phase, while in the second one
the solute from intact cells diffuses first to broken cells and then to
the fluid phase.
For vegetable oil extraction, Sovová [17] found that extraction
curves are initially linear with a slope close to the value of oil sol-
ubility in CO2 . From the extraction curves the initial slope was
calculated and compared with data of solubility of oil in carbon
dioxide. In a recent study, del Valle et al. [19] proposed a general
equation to predict the solubility of vegetable oils in high-pressure
CO2 (within ±40% of experimental values) based on a function of
solvent density () and absolute temperature (T), since solubility
data of various vegetable oils are similar:
  [9.59−8.45((/910)−1)−23.0((/910)−1)2 ]
csat (g · kg−1 ) = 8.07
910
  1 1
  1 1

× exp −4182 1 − 259 − − (3)
T 313 T 313

Fig. 2. Influence of extraction pressure on corn germ oil extraction yield at constant Del Valle et al. [19] state that Eq. (3) can be applied to sys-
extraction temperature of (a) 40 ◦ C ( 30 MPa;  45 MPa;  53 MPa) (b) 80 ◦ C ( tems pure oil + high-pressure CO2 as well as to oil containing
21 MPa;  45 MPa). The solid lines correspond to the model of Sovová [17]. vegetable substrates, since the initial stages of the extraction
process is typically solubility-controlled. To compare the initial
slopes of the extraction curves, data obtained at a process tem-
perature different from 40 ◦ C (T = 313.15 K) were divided by the
temperature-correction term (TCT) of the general model proposed
by del Valle et al. [19]:
  1 1
  1 1

TCT = exp −4182 1 − 259 − − (4)
T 313 T 313
The corrected (at 40 ◦ C) initial slope values obtained from the
first part of the extraction curves have been plotted in Fig. 4 as a
function of pure CO2 density. In this figure, the prediction of the oil
solubility from the General Model proposed by del Valle et al. [19]
together with the error limit of this model has also been plotted.
As it can be observed, the values of the slope of the first part of the
extraction are within the error limits for solubility of vegetable oil
in CO2 .
Rónyai et al. [7] also observed a linear increase at the begin-
ning of the process in the study of SFE of corn germ with carbon
dioxide–ethyl alcohol mixtures. They also explained that the sol-
Fig. 3. Influence of solvent flow rate on corn germ oil extraction yield at a constant vent is saturated with oil in the first stages of the extraction and
extraction pressure (45 MPa) and temperature (79 ◦ C):  3.9 kg CO2 /h ♦ 9.5 kg CO2 /h. solubility of the corn germ oil can be determined from this linear
The solid lines correspond to the model of Sovová [17]. part. The value obtained by Rónyai et al. [7] has been also plotted
in Fig. 4. In this figure, the solubility values of corn germ oil at 40 ◦ C
obtained by Soares et al. [20] have been also included.
Based on these results, Eqs. (5) and (6) proposed by Sovová [17]
3.2. Modeling of the supercritical fluid extraction
were used to evaluate the first and second part of the extraction
curve, respectively:
In this work, the model proposed by Sovová [17] was used to
describe the experimental extraction curves. This type of model e = qys , for 0 ≤ q ≤ qc (5)
assumes that the solute is regarded as a single pseudo com-
pound. This simplification can lead to some errors since several e = xu [1 − C1 exp(−C2 q)], for q > qc (6)
274 S. Rebolleda et al. / J. of Supercritical Fluids 72 (2012) 270–277

Table 2
Values of the C1 , C2 parameters, qc , estimated grinding efficiency r, solid-phase mass transfer coefficient, ks as and O.F.

Experiment C1 C2 qc r ks as O.F.

R1 0.6049 0.0054 25 0.43 1.9 × 10−5 5.0 × 10−3

R2 0.6728 0.0144 26 0.44 5.0 × 10−5 3.2 × 10−3
R3 0.6546 0.0188 23 0.47 6.6 × 10−5 5.7 × 10−3
R4 0.6586 0.0214 22 0.48 6.2 × 10−5 3.8 × 10−3
R5 0.6419 0.0050 55 0.44 2.0 × 10−5 1.5 × 10−3
R6 0.7478 0.0023 75 0.32 1.0 × 10−5 1.2 × 10−2
R7 0.7234 0.0297 17 0.44 7.1 × 10−5 7.7 × 10−3
R8 0.6509 0.0158 28 0.48 2.3 × 10−5 3.2 × 10−3

solid-phase mass transfer coefficient, ks as , can be estimated from

constants C1 , C2 and co-ordinate qc at the crossing point:
 −C q 
2 c
r = 1 − C1 exp (8)
2

(1 − r)(1 − ε)Q̇ C2
ks as = (9)
Nm

In Eq. (9) solvent flow rate is expressed in kg s−1 . Fitting param-

eters and the values of the objective function along with the
estimated values of the grinding efficiency and solid-phase mass
transfer coefficients are collected in Table 2. The values obtained
for the solid-phase mass transfer coefficients, ks as , are of the same
order than those obtained by Sovová [17] when correlating super-
critical extraction data of almond oil [18]. The estimated grinding
efficiency was similar in all the extraction runs and it can be con-
cluded that the volumetric fraction of broken cells in the corn germ
particles is nearly 0.5. The crossing point, qc , was found to increase
Fig. 4. Corrected (at 40 ◦ C) experimental solubility values of corn germ oil as func-
with a decrease in the solubility value, specially marked at low
tion of pure CO2 density ( experimental data points); (—) prediction of del Valle
et al. [19] General Model; (- - -) error limits of the General Model; solubility data
operating pressure.
from Soares et al. [20] (); slope datum from Rónyai et al. [7] ().
3.3. Quality and stability of corn germ oil
C1 and C2 are adjusting constants, ys is the experimental solubility
The quality of corn germ oil was evaluated in four samples
datum, qc the crossing point and xu is the solute concentration in
extracted with SC-CO2 at a constant pressure near 45.0 MPa and
the untreated solid (kg solute/kg insoluble solid). The concentra-
two different temperatures of 56 and 84 ◦ C (runs 9–12).
tion in the untreated solid, xu , was calculated from the oil content
The physical parameters evaluated (refraction index, density
in the corn germ, being xu = 0.8467. The constants C1 and C2 of
and color) showed no difference between the oils extracted at dif-
the model were obtained by minimizing the root squared mean
ferent temperatures. The average values were a refraction index of
deviation between experimental and calculated yield [21]:
1.472 ± 0.001 at 25 ◦ C, a density of 0.92 ± 0.01 kg/L at 20 ◦ C and a
n 2 color expressed as 100.07, −0.007, −0.012 corresponding to light-
i=1
(eexp − ecalc )/eexp
O.F. = (7) ness (L*), redness (a*) and yellowness (b*), respectively.
n
A total of 8 fatty acids were identified and quantified (Table 3) in
by using the Simplex–Nelder–Mead method. The calculated extrac- the different extracts analyzed. It can be observed that the extrac-
tion curves are plotted in Figs. 1–3. From these figures a good tion temperature does not influence the fatty acid profile where
agreement can be observed between experimental data and model linoleic acid (more than 50%) is the major fatty acid, followed by
correlation. According to Sovová [17], the volumetric fraction of oleic acid and palmitic acid. Vigh et al. [6] reported similar fatty acid
broken cells in the particles, called grinding efficiency, r, and the composition of corn oil obtained by SFE. Table 3 also includes the

Table 3
Fatty acid profile of corn germ oil extracted with SC-CO2 and comparison with crude oil.

Fatty acid R9 56 ◦ C R10 57 ◦ C R11 84 ◦ C R12 84 ◦ C R9–R12 Crude oil

mg/g oil mg/g oil mg/g oil mg/g oil (g/100 g fatty acids) (g/100 g fatty acids) [22]

C16:0 107 ± 1 106 ± 3 105 ± 3 106 ± 1 12.5 ± 0.5 11.1–12.8

C18:0 20 ± 1 20 ± 1 19 ± 1 19 ± 1 2.3 ± 0.2 1.4–2.2
C18:1 n-9 245 ± 1 240 ± 4 234 ± 6 234 ± 1 28.2 ± 0.9 22.6–36.1
C18:1 n-7 5 ± 1 5 ± 1 5 ± 1 5 ± 1 0.6 ± 0.1 –
C18:2 n-6 475 ± 2 467 ± 9 455 ± 13 457 ± 2 54 ± 2 49.0–61.9
C18:3 n-3 8 ± 1 8 ± 1 8 ± 1 8 ± 1 1.0 ± 0.1 0.4–1.6
C20:0 4 ± 1 2 ± 1 3 ± 1 3 ± 1 0.4 ± 0.1 0.0–0.2
C20:1 n-9 3 ± 1 2 ± 1 3 ± 1 3 ± 1 0.3 ± 0.1 –

Saturated fatty acids 131 ± 3 128 ± 5 127 ± 5 128 ± 3 15.2 ± 0.8 12.5–15.2
Monounsaturated fatty acids 253 ± 3 247 ± 6 242 ± 8 242 ± 3 29 ± 1 22.6–36.1
Polyunsaturated fatty acids 483 ± 3 475 ± 10 463 ± 14 465 ± 3 56 ± 2 49.4–63.5
Total fatty acids 867 ± 9 850 ± 21 832 ± 27 835 ± 9 100 ± 4 84.5–114.8
 S. Rebolleda et al. / J. of Supercritical Fluids 72 (2012) 270–277 275

Table 4
Neutral lipids profile of corn germ oil extracted with SC-CO2 .

Neutral lipids % wt in oil

R9 R10 R11 R12

Triacylglycerides (TAG) 95.1 ± 0.8 94.4 ± 0.7 95.3 ± 0.4 95.4 ± 0.4
Free fatty acids (FFA) 0.8 ± 0.2 1.1 ± 0.1 1.1 ± 0.1 1.0 ± 0.1
Sterols (St) 2.3 ± 0.4 2.7 ± 0.2 2.2 ± 0.1 2.2 ± 0.1
Others 1.7 ± 0.2 1.8 ± 0.4 1.5 ± 0.2 1.4 ± 0.2

fatty acid profile, expressed in weight percentage, for crude germ factors affect the oxidative stability of oils, such as fatty acid com-
oil described in literature obtained by pressing, followed in some position, stability of antioxidants and the presence of prooxidant
cases by solvent extraction [22]. It can be concluded from Table 3 compounds (FFA, lipid peroxides, or prooxidant metals) [3,26]. List
that fatty acid composition of SC-oil extracted is typical for corn oil and Friedrich [3] suggested that the absence of phosphatides could
fatty acid profile. cause a decrease in the positive synergistic effects of tocopherols
Oil acidity is an important quality parameter related to the pres- with phospholipids. In contrast, Calvo et al. [27] suggested that oil
ence of free fatty acids (FFA) and other non-lipid acid compounds. instability may be related to the trace amounts of oxygen in the
FFA are mainly generated by a hydrolysis reaction of triacylglyc- CO2 . In this case, oxidation would take place without mass transfer
erides, whereas non-lipid acid compounds, such as acetic acid, may limitation since solvent and oil are in the same phase.
be generated during spoilage of the raw material. Thus, oil acidity The higher values of peroxide content reported in this work
depends on several factors related to oil composition, the extraction for freshly SC-CO2 extracted oils cannot be due to the tocopherol
procedure and the raw material freshness. The acidity index (AI) of content in the SC-CO2 extracted oil since its content (Table 5) is even
the oils extracted at 84 ◦ C (1.6 ± 0.1% oleic acid) is of the same order higher than the values reported for hexane prepress extracted wet
as the AI of oils extracted at 56 ◦ C (1.3 ± 0.1% oleic acid). These val- corn germ (1000 ␮g/g) [2]. Tocopherols have been described as one
ues are close to the lowest value of AI described in the literature for of the most effective antioxidants present in vegetable oils. So, its
crude corn germ oils obtained by conventional extraction (from 1.5 presence contributes favorably to the conservation and the qual-
to 4.0% oleic acid) [22]. These values agree with the results reported ity of oils. The tocopherol profile determined by HPLC (Table 5)
by List et al. [2] and Christianson et al. [4] who compare the free was qualitatively similar in all the oils extracted; ␥-tocopherol
fatty acid content of corn oil processing by expeller with those oils was found to be the major one followed by ␣-tocopherol and ␤-
resulting of the SFE of corn germ obtaining lower values of acidity tocopherol, along with a small amount of ␦-tocopherol (Fig. 5).
in the latter. SFE followed by fractionation has been proposed to However, a temperature increase seems to increase the yield of
obtain oils with high quality and less acidity [23]. Some fraction- tocopherols (Table 5). This fact can explain the reduction in the
ation experiments (R13–15) carried out in this work are explained peroxide content at the highest extraction temperature studied in
later. this work. The antioxidant activity of the tocopherols is mainly
The sterol content (Table 4) was nearly the same for the two due to their ability to donate their phenolic hydrogens to lipid
temperatures studied (2.2–2.7 wt% of the neutral lipids). List et al. free-radicals [28]. The increase of tocopherol levels, at the highest
[2] reported values slightly lower for the unsaponificable content of extraction temperature, increases the antioxidant level of oils (see
SC-CO2 extracted corn oil, in the range of 1.2–1.3 wt%, being sterols value of DPPH in Table 5). At constant pressure, the solvent power of
the majority. carbon dioxide decreases with increasing temperature because of
One important quality parameter of fat is the oxidation degree. the decreasing density. Based on these results, the tocopherol con-
Lipids oxidation involves three different stages, initiation, propa- tent found at the highest temperature studied in this work could
gation and termination which rate depend on the substrates and be due to an increase of vapor pressure of the tocopherols with
reaction conditions. Peroxide value is a common parameter eval- temperature. In contrast, List et al. [2] obtained SC-CO2 extracted
uated in oils which is related to the primary oxidation products, corn germ oil with less amount of tocopherols in the extracted oil
being indicative of the deterioration level of the oil where lipid when temperature was increased form 70 to 90 ◦ C at a constant
radicals are attacked by oxygen to form lipid hydroperoxides. Tem- pressure of 82.7 MPa (ranged from 1.840 to 1.180 ␮g/g at 70 ◦ C and
perature is one of the factors that could affect the initiation stage 90 ◦ C, respectively). Nevertheless, these authors [2] concluded that
and therefore the oxidation of the oil [24]. However, in this work, the reason for this drop was unclear. No solubility data of toco-
a slightly decrease in the peroxide content of SC oil corn germ pherols in CO2 were found in the literature for the pressure and
obtained at the highest studied extraction temperature can be temperature range used in this work.
observed (Table 5). In any case, in this work, considerably high
values of the peroxide content have been obtained (Table 5), spe- 3.4. Fractionation experiments
cially when comparing with other values reported in literature for
SC-CO2 extraction of dry, freshly milled corn germ, which rarely In order to improve the oil features, an extraction–fractionation
exceed 0.5 meq/kg [25]. Additionally, List and Friedrich [3] noted process in two separators installed in series was proposed. The first
that SC-CO2 extracted seed oils undergo rapid deterioration. Many separator was maintained at 10.0 MPa and 40 ◦ C. The influence of

Table 5
Peroxide value, antioxidant capacity (DPPH) and tocopherol content of SC-CO2 extracted oils (ND: not detected).

R9 56 ◦ C R10 57 ◦ C R11 84 ◦ C R12 84 ◦ C

Peroxide value (meq/kg oil) 25 ± 2 24 ± 1 21 ± 1 20 ± 1

DPPH (mmol BHT/kg oil) 22 ± 3 25 ± 2 34 ± 1 32 ± 3
Tocopherol content (ppm) 1082 ± 14 1090 ± 3 1481 ± 8 1397 ± 17
␣-Tocopherol 71 ± 5 59 ± 1 97 ± 2 86 ± 2
␤-Tocopherol 65 ± 1 63 ± 1 65 ± 1 64 ± 1
␥-Tocopherol 945 ± 8 967 ± 1 1303 ± 4 1232 ± 11
␦-Tocopherol ND ND 17 ± 1 15 ± 3
276 S. Rebolleda et al. / J. of Supercritical Fluids 72 (2012) 270–277

Fig. 5. Chromatogram of tocopherols in the SC-CO2 extracted corn oil.

Fig. 6. Tocopherol concentration of fractionated oils in the first separator (10.0 MPa and 40 ◦ C) at different extraction temperatures (a) and pressures (b).

extraction pressure and temperature on the tocopherol content of extraction pressure of around 49.0 MPa. This result agrees with the
the oils and on their stability to oxidation was evaluated in these values of tocopherol content presented in Table 5 where the effect
experiments (runs 13–15). It was observed that the fraction col- of extraction temperature on the oil quality was studied. In contrast,
lected in the first separator was mostly oil whereas the fraction Wilp and Eggers [5] found that an increase in extraction tempera-
recovered in the second one was mostly an aqueous emulsion. ture from 50 ◦ C to 80 ◦ C at constant extraction pressure of 50 MPa
Furthermore, it was observed that the induction time determined results in a decrease in tocopherols concentration in the first sep-
by the rancimat test in the oil fraction recovered in separator 1 arator of more than 30%. At constant temperature, around 85 ◦ C,
(1.9 ± 0.3 h) was significantly higher than the induction time deter- when the extraction pressure is increased from 26 MPa to 48 MPa,
mined in the fraction recovered in separator 2 (0.5 ± 0.3 h). The the total tocopherols concentration found in the oil fraction recov-
higher stability found in the oil fraction recovered in separator 1 ered in the first separator was only slightly increased (less than
may be explained considering that both water and free fatty acids by 10%). Similar results were obtained by Wilp and Eggers [5] for
with a high tendency to oxidation, are mostly removed to separa- a decrease in extraction pressure from 50 MPa to 32 MPa at a con-
tor 2. In any case, the induction time of the oil fraction recovered in stant extraction temperature of 50 ◦ C. Based on the results obtained
the first separator is lower than values reported in literature [26] in this work, it could be concluded that the effect of temperature
for hexane Soxhlet extracts of corn germ (3.91 ± 0.4 h). This fact on the extraction of tocopherols is more noticeable than the effect
was previously explained since SC-CO2 extracted oils suffer rapid of pressure.
deterioration.
Fig. 6 shows the tocopherol concentration found in the oil 4. Conclusions
fraction recovered in the first separator at different extraction
conditions. It can be observed that the total concentration of toco- Supercritical carbon dioxide extraction has been studied
pherols in the oil fraction recovered in separator 1 is almost double as a procedure to obtain oil from milled corn germ. Extrac-
when temperature increased from 35 ◦ C to 86 ◦ C at a constant tion experiments have been performed at different extraction
 S. Rebolleda et al. / J. of Supercritical Fluids 72 (2012) 270–277 277

pressure (20.0–53.0 MPa), temperature (35–86 ◦ C) and solvent flow [7] E. Rónyai, B. Simándi, S. Tömösközi, A. Deák, L. Vigh, Z. Weinbrenner, Supercrit-
rate (4–9 kg CO2 /h). The extraction curves obtained indicate that ical fluid extraction of corn germ with carbon dioxide–ethyl alcohol mixture,
The Journal of Supercritical Fluids 14 (1998) 75–81.
the extraction process may be controlled by the solubility of the oil [8] J. Friedrich, E. Pryde, Supercritical CO2 extraction of lipid-bearing materials and
in SC-CO2 in the first stages of the extraction. The external mass characterization of the products, Journal of the American Oil Chemists’ Society
transfer resistance has been found not to be important. 61 (1984) 223–228.
[9] D.D. Christianson, J.P. Friedrich, Production of food grade corn germ product by
Extraction temperature seems to affect the oil quality yield- supercritical fluid extraction, US 4495207, January 22, 1985.
ing oils with higher tocopherol content and therefore higher [10] E.M. Vaquero, S. Beltrán, M.T. Sanz, Extraction of fat from pigskin with super-
antioxidant activity and lower oxidation level when the extraction critical carbon dioxide, The Journal of Supercritical Fluids 37 (2006) 142–150.
[11] CIE, Colorimetry, 2nd ed., Central Bureau of the Commission Internationale de
temperature is increased.
L’Eclairage, Vienna, 1986.
SC-CO2 extraction of corn germ permits to obtain oil with good [12] AOAC, Fatty acids in encapsulated fish oils and fish oil methyl and ethyl esters.
quality with lower values of acidity than for oils obtained by con- Gas chromatographic method, Association of Official Analytical Chemists
41.1.30 (1995).
ventional extraction. A high antioxidant capacity was also found
[13] A. Schaefer, T. Küchler, T.J. Simat, H. Steinhart, Migration of lubricants from
for SC-CO2 corn germ oil. Based on these results, a simpler further food packagings: screening for lipid classes and quantitative estimation using
processing than the traditional one would probably be necessary normal-phase liquid chromatographic separation with evaporative light scat-
for SC-CO2 corn germ oil. On line fractionation in two separators tering detection, Journal of Chromatography A 1017 (2003) 107–116.
[14] IUPAC, Determination of tocopherols and tocotrienols in vegetable oils and fats
installed in series allows separating the co-extracted water and by high performance liquid chromatography, IUPAC (1988).
improves the oil stability against oxidation. [15] W. Brand-Williams, M.E. Cuvelier, C. Berset, Use of a free radical method to
Moreover, the defatted corn germ meal obtained after removing evaluate antioxidant activity, Food Science and Technology 28 (1995) 25–30.
[16] U. Salgın, O. Döker, A. Calımlı, Extraction of sunflower oil with supercritical
the oil with SC-CO2 could be used in food formulation due to its CO2 : experiments and modeling, The Journal of Supercritical Fluids 38 (2006)
good stability and protein quality, however more studies should 326–331.
be done to examine the functional properties of the defatted corn [17] H. Sovová, Mathematical model for supercritical fluid extraction of natural
products and extraction curve evaluation, The Journal of Supercritical Fluids
germ flour and protein isolates. 33 (2005) 35–52.
[18] E. Reverchon, C. Marrone, Modeling and simulation of the supercritical CO2
Acknowledgments extraction of vegetable oils, The Journal of Supercritical Fluids 19 (2001)
161–175.
[19] J.M. del Valle, J.C. de la Fuente, E. Uquiche, A refined equation for predicting the
S.R. acknowledges the PIRTU program of the JCyL Education Min- solubility of vegetable oils in high-pressure CO2 , The Journal of Supercritical
istry and the European Social Fund. The authors acknowledge G. Fluids 67 (2012) 60–70.
[20] B.M.C. Soares, F.M.C. Gamarra, L.C. Paviani, L.A.G. Gonçalves, F.A. Cabral,
Sanjuan for useful comments.
Solubility of triacylglycerols in supercritical carbon dioxide, The Journal of
Supercritical Fluids 43 (2007) 25–31.
References [21] E. Langa, G.D. Porta, A.M.F. Palavra, J.S. Urieta, A.M. Mainar, Supercritical fluid
extraction of Spanish sage essential oil: optimization of the process parameters
and modelling, The Journal of Supercritical Fluids 49 (2009) 174–181.
[1] F.D. Gunstone, Vegetable Oils in Food Technology: Composition, Properties and
[22] L.A. Johnson, Corn: the mayor cereal of the Americas, in: K. Kulp, J.G. Ponte
Uses, Blackwell, Oxford, 2002.
(Eds.), Handbook of Cereal Science and Technology, Marcel Dekker, New York,
[2] G.R. List, J.P. Friedrich, D.D. Christianson, Properties and processing of corn
2004.
oils obtained by extraction with supercritical carbon dioxide, Journal of the
[23] N. Rubio-Rodríguez, S.M. de Diego, S. Beltrán, I. Jaime, M.T. Sanz, J. Rovira,
American Oil Chemists’ Society 61 (1984) 1849–1851.
Supercritical fluid extraction of fish oil from fish by-products: a comparison
[3] G.R. List, J.P. Friedrich, Oxidative stability of seed oils extracted with supercrit-
with other extraction methods, Journal of Food Engineering 109 (2012) 238–
ical carbon dioxide, Journal of the American Oil Chemists’ Society 66 (1989)
248.
98–101.
[24] K. Anna, Lipid oxidation in food systems, in: A.K. Zdzisław, E. Sikorski (Eds.),
[4] D.D. Christianson, J.P. Friedrich, G.R. List, K. Warner, E.B. Bagley, A.C.
Chemical and Functional Properties of Food Lipids, CRC Press, 2002.
Stringfellow, G.E. Inglett, Supercritical fluid extraction of dry-milled
[25] G.R. List, J.P. Friedrich, J.W. King, Supercritical CO2 extraction and processing of
corn germ with carbon dioxide, Journal of Food Science 49 (1984)
oilseeds, Oil Mill Gazetteer (December) (1989) 28–34.
229–232.
[26] J.K. Winkler-Moser, L. Breyer, Composition and oxidative stability of crude oil
[5] V.C. Wilp, R. Eggers, Hochdruckextraktion mit mehrstufiger fraktion-
extracts of corn germ and distillers grains, Industrial Crops and Products 33
ierender Separation zur schonenden Gewinnung von Keimölen mit
(2011) 572–578.
hochverdichtetem Kohlendioxid, Fat Science Technology 93 (1991)
[27] L. Calvo, M. Cocero, J. Díez, Oxidative stability of sunflower oil extracted with
348–352.
supercritical carbon dioxide, Journal of the American Oil Chemists’ Society 71
[6] L. Vigh, B. Simandi, A. Deak, Optimization of supercritical fluid extraction of corn
(1994) 1251–1254.
germ oil in a multipurpose extractor, in: T.H. Applewhite (Ed.), Proceedings
[28] A. Kamal-Eldin, L.A. Appelqvist, The chemistry and antioxidant properties of
of the World Conference on Oilseed Technology and Utilization, AOCS Press,
tocopherols and tocotrienols, Lipids 31 (1996) 671–701.
Champaigne, IL, 1993, p. 433.