Extraction of corn germ oil with supercritical CO₂ and cosolvents

Abstract

The aim of this work was to investigate corn germ oil extraction using supercritical CO₂ and cosolvents addition (hexane, acetone and ethanol). The effects of temperature (45–85 °C) and pressure (15–25 MPa) on the extract yield were evaluated for the tests conducted only with supercritical CO₂. The addition of cosolvents to supercritical CO₂ was also examined at 25 MPa and 60 °C. The conventional Soxhlet extraction with different organic solvents was also performed for comparison purposes. The results of extraction with supercritical fluid showed that the yields increased with pressure at each temperature, but decreased with temperature increase. Mathematical modeling was applied to describe extraction curves, with very good fits. The addition of cosolvents led to higher yield, with a maximum yield of 13.81% using ethanol. The analysis of fatty acids profile did not present significant differences among the evaluated methods. On the other hand, the antioxidant activity of the extracts obtained by supercritical CO₂ extraction was higher than the ones verified for the extracts collected after conventional Soxhlet extraction. Therefore, the use of supercritical CO₂ extraction could be an interesting way to preserve antioxidant properties of this oil in order to use it for pharmaceutical purposes.

Introduction

Corn germ oil is used for cooking and salad oils, margarines, spreads and in animal nutrition due to its beneficial effects against DNA damage, on blood pressure, platelet aggregation and diabetes. It presents high essential fatty acid content, linoleic acid and low levels of linolenic acid that are desirable in vegetable oils (Gunstone 2011).

The industrial process to produce corn germ oil involves extraction using large amounts of hexane or others organic solvents. Due to the toxicity of these organic solvents for food industry, an additional process of evaporation-concentration is required for solvent removal. This conventional extraction with organic solvent takes long time and uses high temperature which can degrade some thermolabile compounds (Rebolleda et al. 2012; Pessoa et al. 2015). Supercritical fluid extraction (SFE) has been studied as a feasible method to oil removal from seeds, since it presents advantages such as high solvation power, use of solvents generally recognized as safe (GRAS), lower extraction times and contamination-free (Asep et al. 2016; da Silva et al. 2016). This methodology has been widely used in the extraction of various vegetable oils, including soybean (Jokić et al. 2012), palm (Dal Prá et al. 2016), pequi (Pessoa et al. 2015), canola (Pederssetti et al. 2011) and candeia (Souza et al. 2008). The most commonly used supercritical fluid is carbon dioxide due to its harmless nature, non-polluting and no toxic character. Besides that, the low critical temperature and pressure of CO₂ (T_c = 31 °C and P_c = 7.38 MPa) enable to operate the extraction process at low temperature (below 80 °C) and moderate pressure (10–45 MPa) which may be an ideal condition for thermo labile compounds extraction (Salea et al. 2014).

Even at high densities, CO₂ has a limited ability to dissolve compounds with high polarity. The addition of cosolvents–organic solvents combined with CO₂—has been used to improve the extraction efficiency by raising the solute solubility and modifying the selectivity. The cosolvent can change characteristics such as polarity and specific interactions with the solute through hydrogen bonds or the active sites of the solid matrix (Dalmolin et al. 2010).

In the literature, few studies of corn germ oil extraction at supercritical conditions with the use of CO₂ and cosolvents were reported. Christianson et al. (1984) investigated the supercritical extraction of corn germ oil at pressures of 34–55 MPa and constant temperature of 50 °C. Experimental assays resulted in higher quality oil with a maximum yield of 23.4% (g/g) of oil at the highest rated pressure value (55 MPa) and better storage characteristics than the oil obtained by extraction with hexane, due to its light color and taste. However, Christianson et al. (1984) observed that high pressures causes proteins denaturation, including of oxidative enzymes.

The quality of the corn germ functional properties was evaluated by Ronyai et al. (1998) with the use of ethanol–water mixture as cosolvent in supercritical CO₂ extraction. The experiments were conducted at constant pressure of 30 MPa and temperature of 42 °C and with the ethanol–water mixture concentration in the CO₂ varying from 0 to 10% by weight. The authors observed that using CO₂ and the addition of 7.5% (wt %) of the ethanol–water mixture, the yield obtained was greater than 50% (g/g). They concluded that the functional properties of the oil obtained from corn germ (sensory properties, emulsifiers and fatty acid composition) presented satisfactory results for use in the food industry.

Rebolleda et al. (2012) analyzed the influence of temperature (35–86 °C), pressure (20–53 MPa) and CO₂ mass flow rate (4–9 kg/h) on yield and quality of corn germ oil obtained with supercritical CO₂ extraction. The oil presented antioxidant capacity with maximum value of 34 mol BHT/kg (mol of butylated hydroxytoluene per kilogram), at a temperature of 84 °C, pressure of 44 MPa and CO₂ flow rate of 6 kg/h. The effect of temperature and pressure on the extraction showed that high temperatures and pressures increased the extraction yields. At 45.0 MPa and temperatures of 40 and 79 °C, yields of 0.65 and 0.80% (kg/kg), respectively, were obtained. At a fixed temperature of 85 °C and pressures of 45 and 53 MPa, the yield values were 0.45% and 0.65% (kg/kg), respectively. The solvent mass flow rate had no influence on the extraction yield under the conditions used (4–9 kg/h).

The objective of this work was the extraction of corn germ oil using supercritical CO₂ as solvent and cosolvents addition. The experiments were performed at the temperature range of 45–85 °C, pressure range of 15–25 MPa, and fixed flow rate of 3 mL CO₂/min in the extraction with supercritical CO₂. In the extractions with cosolvents addition, three different compounds were tested (hexane, acetone and ethanol) at a fixed pressure of 25 MPa and 60 °C. At the end of the extraction processes, the quality and the extracts composition were analyzed and the results were compared with the conventional extraction method in Soxhlet.

Materials and methods

Materials

Corn germ was provided by Caramuru Alimentos S.A., Itumbiara-Goiás, Brazil. The properties of the full-fat, dry-milled germ by weight were 11.26 ± 0.09% humidity, 11.38 ± 0.13% protein, 3.28 ± 0.05% ash, and 7.67 ± 0.47% fiber. Particles with size between 16 and 20 mesh (#16 = 1.19 mm and #35 = 0.5 mm) were used in this study. The material was used for all the SFE extractions and Soxhlet extraction.

Extraction methods

Organic solvent (Soxhlet)

The organic solvent extraction was performed in a Soxhlet apparatus using the Adolfo Lutz Institute (2008) methodology. Adolfo Lutz Institute in a Brazilian official laboratory from the Health Ministry. The solvents used were hexane (Synth, 98.5%, Brazil), ethanol (Anidrol, 99.5%, Brazil) and acetone (Anidrol, 99.5%, Brazil). The yields obtained for each solvent extraction were expressed relative to the initial weight sample.

Supercritical fluid extraction with CO₂ and with cosolvents addition

The experiments were performed in a laboratory scale unit, which consisted of a CO₂ cylinder (WHITE MARTINS S.A., 99.5% of purity, Brazil), one syringe pump (ISCO, Model 500D, U.S.A), one extractor and two thermostatic baths. The extractor has an internal capacity of 63 cm³, approximately (diameter of 1.98 cm and height 20.5 cm) which was loaded with approximately 25 g of sample in each experiment. One of the thermostatic baths (THERMO SCIENTIFIC, A-10, U.S.A) was used in order to cool the solvent before entering in the pump and the other one (TECNAL, TE-184, Brazil) was used to keep the extractor heated to the desired temperature. A pre-extraction time of 30 min was used and then the extraction was initiated. The weight of the collection flask was measured every 10 min in the first hour and every 30 min in the rest of the hours. The weight was determined with a Shimadzu analytical balance (model ATX224, Brazil) that has a resolution of ± 0.1 mg. This unit and the procedure have been already described in detail by Souza et al. (2008).

In order to evaluate the influence of temperature and pressure on the supercritical fluid extraction yield using CO₂, 9 experiments were carried out. The influence of density on temperature and pressure was studied. CO₂ densities were consulted in literature (NIST 2017). The process variables evaluated were extraction pressure (15–25 MPa) and temperature (45–85 °C).

The experiments of supercritical fluid extraction with the use of CO₂ and cosolvents (hexane, ethanol and acetone) were carried out in the same laboratory scale unit as the experiments using only CO₂. The methodology was described in the supercritical fluid extraction section, except by the addition of cosolvents with an isocratic pump simultaneously with the CO₂ input during the extraction process.

The extractions were performed at 60 °C, flow rate of 0.10 mL/min of cosolvent, fixed pressure of 25 MPa, CO₂ flow rate of 3 mL/min for 4 h. The sample was weighed only at the end of the process, after rotavaporation and drying of the sample in an oven.

The yields were calculated by Eq. (1) and adjusted in a quadratic model using regression analysis and p value lower than 0.05 was considered as statistically significant.

$$Yield=\frac{m_{\mathit{ext}}}{m_{\mathit{dryext}}}\times 100$$ 1

In which:

• $m_{\mathit{ext}}$ = total mass of extracted oil (g);

• $m_{\mathit{dryext}}$ = mass of dry extract (corn germ) (g).

Corn germ oil characterization

Fatty acids profile analysis

In order to determine the fatty acids, 60 mg of corn germ oil samples were derivatized. To each sample, 2 mL of the previously prepared KOH (Vetec, 85%, Brazil)/Methanol (Dinâmica, 99.8%, Brazil) solution was added. The mixture was homogenized for 5 min in a vortex and 5 mL of heptane (Synth, 99%, Brazil) was added. After that, the upper phase formed by the two-phase separation was transferred to the vials where the fatty acid quantification of corn germ oil was carried out. The derivatized samples were analyzed in a gas chromatograph (Shimadzu, model GC-2010) equipped with FID detector following the AOCS method Ce 2-66 (AOCS 2009). The capillary column was RTX-Wax (0.32 mm i.d. × 30 m, film thickness 0.25 µm-Restek) and helium was used as the carrier gas, with a split ratio of 1:50. The temperatures of the column and the injector were 210 and 250 °C, respectively. The identification and quantification of the fatty acids peaks were performed by comparison of the retention time of each compound in the sample with the internal standard. The standards of palmitic, stearic, oleic, linoleic, and linolenic fatty acids were used in the quantitative analysis of corn germ oil (Rade et al. 2015).

Mathematical modeling

The overall extraction curves (OEC) that presented the highest yields in the supercritical fluid extraction with CO₂ were fitted with spline (Meireles 2008; Santana et al. 2017) and logistic models (Martínez et al. 2003).

Spline model subdivide the OEC spline in three straight lines. Spline model could be associated to the broken and intact cell model (Sovová 1994) since the experimental and modeled curves displayed the three regions: constant extraction rate (CER), falling extraction rate (FER), and diffusion controlled (DC). The equation of model can be expressed by Eqs. (2)–(4):

$$\text{when}t\le t_{\mathit{CER}}\to m_{\mathit{extracted}}=b_0{m}_{\mathit{Alim}}+Q_{\mathit{sol}}{b}_{1}t$$ 2

$$\begin{array}{cc}& \text{when}{t}_{\mathit{CER}}<t\le t_{\mathit{FER}}\to m_{\mathit{extracted}}=b_0{m}_{\mathit{Alim}}\hfill \\ \hfill & +Q_{\mathit{sol}}\left[b_{1}t+b_2\left(t-t_{\mathit{CER}}\right)\right]\hfill \end{array}$$ 3

$$\begin{array}{cc}& \text{when}t>t_{\mathit{FER}}\to m_{\mathit{extracted}}\hfill \\ \hfill & =b_0{m}_{\mathit{Alim}}+Q_{\mathit{sol}}\left[b_{1}t+b_2\left(t-t_{\mathit{CER}}\right)+b_3\left(t-t_{\mathit{FER}}\right)\right]\hfill \end{array}$$ 4

CER, FER, and DC periods are delimited by the time ending periods of CER (${\text{t}}_{\text{CER}}$) and FER regions (${\text{t}}_{\text{FER}}$). b₀, b₁ and b₂ are parameters. ${\text{b}}_1$ corresponds to the mass ratio of extract in the supercritical phase at the bed outlet at CER step ($Y_{\mathit{CER}}$). ${\text{m}}_{\text{Alim}}$ is the initial mass on the extractor vessel, and ${\text{Q}}_{\mathrm{sol}}$ is the solvent mass flow. The product ${\text{Q}}_{\mathrm{sol}}{\text{b}}_1$ values are equal to ${\text{M}}_{\mathrm{CER}}$ values, which are the extraction rates for the CER period.

The model of Martínez et al. (2003) assumes that the extract is a mixture of compounds. The equation of model can be expressed by Eq. (5):

$$m_{\mathit{extracted}}=\frac{m_0}{exp\left(C{t}_m\right)}\left\{\frac{1+exp\left(C{t}_m\right)}{1+exp\left[C\left(t_m-t\right)\right]}-1\right\}$$ 5

The parameter $t_m$ matches to the time of maximum extraction rate and parameter $C$ has no physical meaning. Logistic model depends on the ${\text{m}}_0$ value which is the initial solute mass. This value was considered as an adjustable parameter, as recommended by Martínez et al. (2003) when these data are not determined experimentally. To use this model, the extract was considered a pseudo-pure substance.

The modeling of spline model was carried out using an algorithm implemented in MS Excel, which was recently disclosed by Santana et al. (2017). Parameters of the logistic model were calculated by non-linear regression using Statistica 10^® software. We used simplex least square fitting procedure with convergence criterion of 10⁻⁶.

Determination of antioxidant capacity: DPPH^· Assay

The antioxidant capacity was determined using the stable radical DPPH^· (2,2-diphenyl-1-picrylhydrazyl) based on the methodology of Brand-Williams et al. (1995) modified. DPPH^· (Sigma-Aldrich) was used as standard to evaluate free radical scavenging capacity of corn germ oil. In dark environment, a 0.1 mL aliquot of extract was added to 3.9 mL of DPPH^· and homogenized. This procedure was performed at different extract concentrations. Absorbance was measured at 515 nm in duplicate. Results were expressed into the extract concentration required to reach 50% of the percentage antioxidant activity, EC₅₀.

Results and discussion

Extraction of corn germ oil processes

Extraction with supercritical CO₂: temperature and pressure effects

Table 1 shows the results of the extraction yield of corn germ oil in SFE with CO₂ after the end process. At constant pressure, there was a decrease in the yield of the supercritical extraction with the increase of temperature (runs 1 and 2 at 16.4 MPa, runs 3 and 4 at 23.5 MPa and runs 5 and 6 at 20.0 MPa).

Table 1.

Yield of CO₂ supercritical extraction in different temperature and pressure conditions

Run	Temperature (°C)	Pressure (MPa)	CO2 density (g/mL)	Yield (%) (Y)
1	52	16.4	0.7146	5.00
2	80	16.4	0.4839	1.90
3	52	23.5	0.8112	6.83
4	80	23.5	0.6633	5.68
5	45	20.0	0.8127	6.71
6	85	20.0	0.5627	3.39
7	65	15.0	0.5537	2.34
8	65	25.0	0.7619	6.41
9	65	20.0	0.6917	5.86

Salgin et al. (2006), while studying the supercritical CO₂ extraction of sunflower oil, verified that a temperature increase caused a significant decrease in the yield and solubility. For instance, at 20.0 MPa and temperatures of 40 and 80 °C, yields of 0.80 and 0.10% (kg/kg), respectively, were obtained. These results can be explained by the crossover effect, which the temperature increasing at constant pressure leads to the decreasing of CO₂ density as well as the increasing of the solute vapor pressure (Jokić et al. 2012). The decrease of CO₂ density is significative at pressures close to the critical point, resulting in a drop in solubility. Özkal et al. (2005) observed the same behavior in solubility with temperature rise in the supercritical CO₂ extraction of apricot seeds at 15 MPa, due to the crossover effect. They determined yields of 1.10, 0.90, and 0.2% (mg/g) for temperatures of 40, 50 and 60 °C, respectively. Rebolleda et al. (2012) studied the supercritical extraction of corn germ oil with CO₂, and verified a contrary effect in the influence of temperature. However, the constant pressure used in the extraction was 45 MPa, higher than the pressure used in the present work.

The effect of pressure on the yield of extracted oil at supercritical conditions was also evaluated. According to Table 1, it was noted that the increase in pressure at the same temperature raised the yield of the oil at constant temperature. Runs 2 and 4 at 80 °C and runs 7, 8 and 9 at 65 °C showed that the increase of pressure caused an increase of extraction yield. Rebolleda et al. (2012), Pederssetti et al. (2011), Nimet et al. (2011) and Subroto et al. (2017) studied the supercritical CO₂ extraction of corn germ, canola, sunflower and candlenut oils, respectively, verified the same effect, mainly at pressures above 20 MPa.

In the present study, we also observe temperature and pressure effects similar to the mentioned authors (Özkal et al. 2005; Salgin et al. 2006; Pederssetti et al. 2011; Nimet et al. 2011; Rebolleda et al. 2012; Subroto et al. 2017). Then, we can imply that SFE extraction of oil from corn gern has remarked influence of solvent solubility and density, since this physical properties are direct related with crossover effect. In experimental range evaluted, we can notice a linear correlation between yields and CO₂ density, as expressed in Fig. 1. For further experiments, this correlation, that achieved 0.93 of correlation coeficiente (R²), can be useful to estimated expected yields.

Fig. 1.

Influence of CO₂ density (ρ) on corn germ extraction yield. The solid line represents the linear fit (R² = 0.93)

Modeling of the supercritical fluid extraction

The OEC that presented higher yields in the supercritical fluid extraction with CO₂ (runs 3, 5 and 8) were fitted with spline and logistic models. Table 2 shows the determination coefficients (R²) and model constants for kinetic models fitted by regression to the experimental data. It is possible to notice that the models described the extraction kinetics very well, given the high R² values. Figure 2 illustrates the goodness of fit, representing mass extracted as a function of time of extraction (t).

Table 2.

Correlations coefficients (R²) and model constants for spline kinetic model fitted to the experimental data

Model	Value	Unit	Run 3	Run 5	Run 8
			52 °C 23.5 MPa	45 °C 20.0 MPa	65 °C 25.0 MPa
Spline	$R^2$	(–)	0.9979	0.9987	0.9991
	$b_0$	(–)	1.014E−03	− 1.190E−03	− 9.372E−04
	$b_1=Y_{\mathit{CER}}$	(g g−1)	1.192E−02	7.264E−03	1.070E−02
	$b_2$	(g g−1)	− 5.569E−03	− 3.613E−03	− 6.077E−03
	$b_3$	(g g−1)	− 5.596E−03	− 2.994E−03	− 3.900E−03
	$t_{\mathit{CER}}$	(min)	25.00	45.00	34.00
	$t_{\mathit{FER}}$	(min)	75.00	135.00	102.00
	$M_{\mathit{CER}}$	(g min−1)	2.896E−02	1.765E−02	2.439E−02
Logistic	$R^2$	(–)	0.9892	0.9952	0.9888
	$m_0$	(g)	1.6722	1.7039	1.6171
	$C$	(min−1)	0.0444	0.0242	0.0364
	$t_m$	(min)	19.96	30.85	24.27

Fig. 2.

Experimental and modeled kinetic curves for the extraction of corn germ oil with CO₂ as solvent: run 3—52 °C and 23.5 MPa, run 5—45 °C and 20.0 MPa, and run 8—65 °C and 25.0 MPa

In the studied kinetics, CER and FER steps, evidenced by their respective time-span periods, varied between 25–45 min and 75–135 min. The optimum conditions provided the highest extraction rate for the constant extraction period (M_CER) and mass ratio of extract in the supercritical phase at the bed outlet (Y_CER) values.

Regarding to logistic model, $m_0$ values estimated are close of mass extracted at DC period, which is expected and indicates good representation of physical problem. $t_m$ values were smaller than $t_{\mathit{CER}}$, which shows that the higher mass transfer rate occurs at CER period, in other words, in the region when OEC was higher angular coefficient.

Extraction with supercritical CO₂ and cosolvents addition

The supercritical CO₂ extraction with the addition of cosolventes at 60 °C led to the following yields: 11.01% hexane, 13.54% acetone and 13.81% ethanol. Based on duplicate experiments, the maximum uncertainties values were ± 0.35 g. It is observed that the use of hexane as cosolvent led to higher yields than extraction with pure supercritical CO₂. However, this extraction showed the smallest increase among the three cosolvents used. This may be justified by the fact that hexane is non-polar as well as CO₂, acting the same way in the extraction, while acetone and ethanol are polar and can similarly extract other compounds that have not yet been extracted. Besides that, ethanol is the most used cosolvent because it is inoffensive for human health and is environmentally friendly. The best oil yield achieved using ethanol as cosolvent is probably due to the ability of this solvent in forming hydrogen bonds, differently from acetone, which is an aprotic solvent (Andressa et al. 2017).

The addition of ethanol to the CO₂ provides an increase of the yield due to the increase of oil solubility. In order to compare the oil yield of extractions under supercritical conditions with the use of CO₂ and cosolvents, extractions with organic solvents were carried out in a Soxhlet apparatus. The highest oil yields were obtained with the use of ethanol (35.70%), followed by hexane (30.24%) and acetone (22.50%). Soxhlet extraction with ethanol presented the highest yield probably due to the polar nature of ethanol, being able to extract not only polar compounds, which do not dissolve in hexane, but also non-polar components.

The contact between the solvent and the substrate in Soxhlet extraction is different from that in supercritical extraction. In the Soxhlet, the solvent is heated to reflux, the steam passes through a “by-pass” to reach the condenser. The solvent drips down into the system and diffuses into the substrate matrix. The substances to be extracted are consequently dissolved in the solvent and this mixture of solvent and extracted substances diffuses from the matrix of the substrate back into the solvent. Once the mixture level in the extractor reaches the top of the siphon arm, the solvent and extract return to the lower flask. Therefore, this is a batch extraction process. In the supercritical extraction with the use of CO₂ and also of cosolvents, once CO₂ diffuses into the substrate matrix and solubilizes the components to be extracted, then the mixture of components and CO₂ diffuses from the substrate matrix and flows with the volume of the fluid phase. Therefore, supercritical extraction is a continuous process, involving transport phenomena different from the Soxhlet extraction, which provides greater process selectivity and, consequently, extraction of compounds of greater interest (Zhao and Zhang 2014).

The higher yields with the use of organic solvents in Soxhlet were obtained for a long extraction time, about 8 h and a large amount of solvent is used, which can cause thermal decomposition of the compounds and small selectivity (Luque de Castro and Priego-Capote 2010). On the other hand, the extractions under supercritical conditions were carried out in a shorter extraction time, of a maximum of 4 h (time dependent on the exhaustion of the extracted sample), which may lead to less oil degradation. This technique uses a solvent recognized as safe for human health and the environment. Moreover, the extract is preserved from contact with light, where reactions could take place causing oil oxidation; the solvent removal and recovery occurs easily with the depressurizing of the system.

Physical oil properties

Fatty acids analysis

The fatty acids analysis of corn germ oil was performed in the following extractions: with supercritical CO₂ in some points (runs 1, 2 and 3) and with the use of cosolvents and CO₂ at 60 °C and 0.10 mL/min (Table 3).

Table 3.

Fatty acids profile obtained by supercritical fluid extractions of corn germ oil with the use of CO₂, with the use of organic solvents and with cosolvents addition (ethanol = E; hexane = H and acetone = A)

Extraction method	Fatty acid (% g/goil)
	Run/solvent	Estearic C18:0	Palmitic C16:0	Oleic C18:1	Linoleic C18:2	Linolenic C18:3
Supercritical CO2	1	1.5 ± 0.3	13.8 ± 0.4	34.8 ± 0.3	47.5 ± 0.6	0.7 ± 0.0
	2	1.5 ± 0.0	12.3 ± 0.1	35.8 ± 0.1	48.7 ± 0.0	0.8 ± 0.0
	3	1.7 ± 0.1	11.7 ± 0.7	37.0 ± 0.3	48.7 ± 0.2	0.7 ± 0.0
Soxhlet	E	1.2 ± 0.1	13.7 ± 0.5	34.6 ± 0.5	49.2 ± 0.0	1.2 ± 0.1
	H	1.2 ± 0.1	11.7 ± 0.7	36.5 ± 0.5	49.3 ± 0.8	1.1 ± 0.4
	A	1.3 ± 0.2	13.8 ± 0.7	37.0 ± 0.7	47.7 ± 0.8	0.5 ± 0.3
Cossolvents (0.10 mL/min and 60 °C)	E	1.2 ± 0.0	12.8 ± 0.4	37.2 ± 0.0	47.6 ± 0.2	0.9 ± 0.1
	H	1.2 ± 0.0	11.4 ± 1.1	37.3 ± 0.0	49.0 ± 1.0	1.0 ± 0.0
	A	1.6 ± 0.5	13.1 ± 2.7	35.7 ± 1.7	48.7 ± 1.7	1.7 ± 0.1

The fatty acids found in highest proportion were the linoleic and oleic unsaturated acids, respectively, for all types of the extraction performed, whereas the saturated fatty acids, palmitic and estearic were found in less quantity. In addition, the results obtained in the quantitative analysis of fatty acids for corn germ oil with the use of different extraction methods showed no significant variation of the fatty acids contents at the conditions or solvents used. Other results published in literature indicated that hexane (Rai et al. 2016), ethanol (Carvalho et al. 2012) and acetone (Dabrowski et al. 2018) could not change lipid profile of the extracts by Sohxlet or as co-solvents in SFE-extractions.

For extractions with supercritical CO₂, the same result in highest proportion of linoleic and oleic unsaturated acids was observed by Rebolleda et al. (2012) for corn germ extraction and also for others vegetable oils, such as canola oil (Pederssetti et al. 2011) and sunflower oil (Nimet et al. 2011). Comparing the results obtained by Rebolleda et al. (2012), the oleic acid profile was significantly higher in this study. The authors obtained values between 23.4 and 24.5% g/g_oil and the linoleic acid composition was almost the same (45.5–47.5% g/g_oil). The linolenic acid composition determined in this study was about 8 times smaller than those found by Rebolleda et al. (2012). These results demonstrate the excellent corn germ oil composition and quality in terms of fatty acids, considering that the tests were realized with the crude oil without subsequent processes to refine the oil, after the extractions.

Determination of antioxidant capacity: DPPH^· Assay

The antioxidant activity of the samples obtained with the use of supercritical CO₂, organic solvents and cosolvents addition were determined with the DPPH^· method. The results for the EC₅₀ obtained from the graph of A.A. % as a function of samples concentrations (µg/mL) are shown on Table 4 as the extracts classification, according to Garmus et al. (2015).

Table 4.

Antioxidant potential of corn germ oil using the DPPH^· method in the extractions with the use of supercritical CO₂, cosolvents addition and organic solvents by Soxhlet

Extraction method	Conditions and/or solvents	EC50 (µg/mL)
Supercritical CO2	1 (T = 52 °C and P = 16.4 MPa)	8.69 ± 0.09
	2 (T = 80 °C and P = 16.4 MPa)	5.03 ± 0.16
	3 (T = 52 °C and P = 23.5 MPa)	7.93 ± 0.00
Soxhlet	Ethanol	12.95 ± 0.22
	Hexane	16.34 ± 0.73
	Acetone	14.90 ± 0.29
Supercritical CO2 + cosolvents	Ethanol (0.10 mL/min and 60 °C)	12.29 ± 0.03
	Hexane (0.10 mL/min and 60 °C)	17.43 ± 1.89
	Acetone (0.10 mL/min and 60 °C)	14.85 ± 1.35

Table 4 shows that all the extracts presented antioxidant activity, being classified as very active. The lowest EC₅₀ values and, therefore, the best results of antioxidant activity were obtained with the CO₂ supercritical extraction. It demonstrates that the compounds responsible for the antioxidant activity in the corn germ oil have higher affinity for non-polar solvents.

The results of the supercritical extraction with CO₂, were obtained at different values of temperature and pressure, parameters that affect the CO₂ density and the extraction process due to the mass transfer. The pressure and temperature effects showed that the increase of these parameters leads to the decrease of EC₅₀ that is an increase of the concentration of the components with antioxidant potential.

The extracts obtained by Soxhlet and with the cosolvents addition presented similar results. Ethanol has the highest oxidizing potential, followed by acetone and hexane, respectively, in according to Kitzberger et al. (2007) which suggested that the ethanol polar nature favors the antioxidant compounds.

Conclusion

The present work applied experimental planning that proved to be an important tool for the study of the supercritical extraction of corn germ oil with the use of CO₂. The results of extraction with supercritical fluid showed that the corn germ oil yield values increased with pressure at each temperature, but decreased with temperature increase. Some overall extraction curves were fitted by logistic and spline models. Both models were satisfactory able to describe the extraction kinetics. The analysis of quality and composition of corn germ oil demonstrated that the corn germ oil presented excellent profile of fatty acids. The large amount of linoleic acid and small amount of linolenic acid showed that the corn germ oil extracted had a good nutritional value and high antioxidant potential in the temperature and pressure conditions studied. The yields of the extracts obtained with the use of CO₂ and addition of cosolvents (hexane, acetone and ethanol) were higher when compared to the yields presented by the supercritical extraction with CO₂. The organic solvents presented the highest values of oil yield, but a long extraction time and a large amount of solvent were necessary, which may degrade and decrease the selectivity of the compounds of interest extracted.