Assessment of subcritical propane, supercritical CO2 and Soxhlet extraction of oil from sapucaia (Lecythis pisonis) nuts

Assessment of subcritical propane, supercritical CO2 and Soxhlet extraction of oil from sapucaia (Lecythis pisonis) nuts

The Journal of Supercritical Fluids 133 (2018) 122–132 Contents lists available at ScienceDirect The Journal of Supercritical Fluids journal homepag...

1MB Sizes 1 Downloads 139 Views

The Journal of Supercritical Fluids 133 (2018) 122–132

Contents lists available at ScienceDirect

The Journal of Supercritical Fluids journal homepage: www.elsevier.com/locate/supflu

Assessment of subcritical propane, supercritical CO2 and Soxhlet extraction of oil from sapucaia (Lecythis pisonis) nuts

MARK

Gerson Lopes Teixeiraa, Saeed M. Ghazanib, Marcos Lúcio Corazzaa, Alejandro G. Marangonib, ⁎ Rosemary Hoffmann Ribania, a b

Graduation Program of Food Engineering, Federal University of Paraná, Polytechnic Center, Jardim das Americas, Curitiba, Paraná, 81531-980, Brazil University of Guelph, Guelph, ON, N1G 2W1, Canada

G RA P H I C A L AB S T R A C T

A R T I C L E I N F O

A B S T R A C T

Keywords: Sapucaia Subcritical propane Thermal behavior Triacylglycerols Oxidative stability

The extraction of sapucaia (Lecythis pisonis) nut oil (SNO) using subcritical propane (SPE) and supercritical CO2 (with ethanol as co-solvent; scCO2) as solvent was investigated and compared with the conventional (Soxhlet) extraction. Extraction with scCO2 was performed at 333 K and 20 MPa while the SPE extractions were carried out in different conditions to investigate the effects of temperature (293–333 K) and pressure (2–10 MPa) on the oil yield and the chemical compositions of the products. Results show that SPE allowed a fast extraction with a higher yield (46.22%) obtained at 333 K and 10 MPa, representing 93% efficiency compared to Soxhlet. Only temperature had significant (p < 0.05) effect on the extraction yield. SPE yielded the oils with highest values of polyunsaturated fatty acids (∼36%). Stability to oxidation ranges from 6.53 to 11.17 h. The major triacylglycerols present in SNO are OOO, SOO, POO, PLO, and POS.

1. Introduction An effective way to valorize tree species and prevent their extinction consists mainly of some reforestation techniques or the reduction of deforestation. Another solution is to find proper use for their principal products (fruits, leaves, seeds), aimed at the maintenance of a cycle that assures protection and survival for the species. Some Amazonian trees that were not adequately studied and products of which are not available in the market have no application in industry, being neglected



because of the lack of the required research to stimulate their usage. In this context, Lecythis pisonis, a tree which is present in most regions of Brazil [1], is underutilized in terms of the use of its main product, the so-called “sapucaia” nuts. This edible nut presents a high lipid content (51–64%), predominantly linoleic acid [1,2] in a yellowish oil, presenting a characteristic flavor and is also well known for its similarity to Brazil nut (Bertholletia excelsa). Although some studies have been conducted with this raw material, its application and usability are still poorly appreciated, because of a lack of information regarding some

Corresponding author. E-mail address: [email protected] (R.H. Ribani).

http://dx.doi.org/10.1016/j.supflu.2017.10.003 Received 5 September 2017; Received in revised form 28 September 2017; Accepted 2 October 2017 Available online 05 October 2017 0896-8446/ © 2017 Elsevier B.V. All rights reserved.

The Journal of Supercritical Fluids 133 (2018) 122–132

G.L. Teixeira et al.

2.2.1. Classical extraction The oil from sapucaia nut samples was extracted using a Soxhlet apparatus (Vidrolabor®, Labor Quimi, Brazil) for 6 h, using n-hexane as solvent [26]. In order to obtain enough oil samples for all the analysis, ∼25 g of raw material was used in the extraction. Solvent was removed at 316.15 K under reduced pressure using a rotary evaporator (Model 801, Fisatom Ltda., Brazil). The sample was flushed with gaseous N2 before storage. The oil was kept in an amber glass vessel and stored at 277.15 K until further analysis.

other processes of extraction with minimum changes in its characteristics. The use of conventional extraction processes that involve organic solvents for extraction is discouraged due to solvent residues and damage to some components, which are important in the final product, mainly due to the use of high temperatures [3]. On the other hand, new techniques which are as effective as the traditional ones and result in high quality products with no solvent remaining have been promoted and extensively stimulated, such as the use of pressurized fluids under sub- or supercritical condition [3–8]. A widely used technique, which is a so-called “green technology”, is the use of carbon dioxide at supercritical condition (scCO2; T = 304.25 K, P = 7.39 MPa) [6]. However, scCO2 is less effective in extracting lipid portions, because of the low solubility of triacylglycerides [4,9], and the use of high-pressure or cosolvents (such as ethanol) being necessary to improve the extraction yields. On the other hand, compressed or subcritical propane (critical conditions: T = 369.82 K, P = 4.25 MPa) has been investigated and used as a solvent in which triglycerides and fatty acids show high solubility [8,9], providing higher extraction yields, the major benefit being the use of lower pressure and temperature which leads to a high quality product with minimum damage, no degradation of bioactive compounds, shorter extraction times and completely free of residue [4,5,7,10,11], thus presenting some advantages over scCO2 [5]. Propane in compressed conditions has been successfully applied in the extraction of oils from a wide range of vegetable sources such as macaúba pulp [4], passion fruit seed [7], kiwi seed [8], palm [12], sesame seed [11], pequi [13], canola seed [14], red pepper seed [15], crambe seed [16], flaxseed [17], perilla [18], Moringa oleifera [19], Sacha inchi [20], Araucaria angustifolia [21], grape seed [22], and sunflower [23]. Publications in the literature report oil extraction from L. pisonis nuts by cold extraction [2], and also using organic solvent by Soxhlet [1,24,25] and Bligh & Dyer methods [25]. However, there are no data about oil extraction from sapucaia nuts using compressed propane and supercritical CO2 processes. Thus, the aim of the study was to obtain sapucaia nut oil using subcritical propane and supercritical carbon dioxide (with ethanol as co-solvent) techniques and investigate the effects of temperature and pressure in the oil extractions. The Soxhlet technique using n-hexane as solvent was applied to obtain an oil that was used for comparison purposes. In addition, the oil samples obtained (from different process and techniques) were analyzed for their fatty acid composition, oxidative stability, crystallization and melting behavior, and triacylglycerol composition.

2.2.2. Sub- and supercritical fluid extraction procedures The propane and CO2 used in this work were purchased from White Martins S.A. (99.5% purity in liquid phase). The extractions were performed in a bench-scale unit (represented by Fig. 2), described in detail in previous works by our research group [5,8]. Briefly, the experimental setup consists of jacketed-vessel (0.08 m3 inner volume, L = 0.16 mm and Φ = 2.52 × 10−2 m) coupled to a thermostatized bath, a micrometering needle valve to control the flow inside the extractor, a syringetype pump (ISCO, model 500D, Lincoln, NE 68504, USA), and pressure and temperature sensors and transducers. Supercritical CO2 extraction condition was stablished according to previous tests performed in our laboratory by using 1:1 (w/w) ethanol as co-solvent by directly immersing the sample into the alcohol just before confinement in the extractor. Extraction using subcritical propane was performed based on a simple 22 factorial experimental design with a center point (Table 1), aiming to evaluate the effects of pressure and temperature on the extraction yield. The solvent was pumped at a constant flow rate of 2.0 ± 0.2 cm3 min−1 for both fluids used. The oil was collected in an amber glass vessel and its weight was determined at time intervals of 5 min of extraction. The yields were calculated as the ratio of the extracted oil mass to the initial sapucaia nut weight. The analysis of the experimental data, at 95% confidence level, was carried out using the Statistica 10.0 software program (Statsoft™, Inc.). 2.3. Fatty acid composition The fatty acid composition of the different sapucaia oils was determined using GC. Oil samples were directly methylated using methanolic sodium methoxide as described by Christie (1982) [27] and then injected (1-μL aliquot) into a capillary BPX70 column, 60 m × 0.22 mm internal diameter with 0.25 μm film thickness (SGE Inc., Austin, TX, USA). An Agilent 6890-series Gas Chromatograph (Agilent Technologies, Inc., Wilmington, DE, USA) with 7683-series auto-sampler was used to house the column. The oven temperature was programmed to increase from 110 to 230 °C at a rate of 4 °C/min and maintained at 230 °C for 18 min. The injector and detector temperatures were 250 and 255 °C, respectively. Helium was used as the carrier gas at an average velocity of 25 cm/s. Fatty acid composition was expressed as the percentage of the total peak area of all the fatty acids in the oil sample.

2. Material and methods 2.1. Sapucaia (Lecythis pisonis) nut samples Sapucaia nut samples were harvested in a small crop area located in Araguanã, State of Maranhão (Brazil). The nuts were dried in an air circulating oven at 313.15 K for 24 h to remove moisture. The shells were then removed using a stainless-steel nut cracker. Peeled nuts were packed in plastic bags, crushed with a small hammer, and passed through a Tyler sieve system composed of sieves with mash numbers 8, 12 and 24. The nut pieces that passed through sieve #8 and were held in the sieve # 12 were used in the extractions, as shown in the schematic Fig. 1.

2.4. Triacylglycerol composition of sapucaia oils The triacylglycerol (TAG) composition of the sapucaia nut oils was evaluated by using a high performance liquid chromatography (HPLC) system. About 30 mg of oil sample was placed in a 2 mL HPLC vial. The sample was dissolved by adding 600 μL chloroform and 1 mL 60:40 HPLC-grade acetone:acetonitrile solution. TAG composition of sapucaia oil was obtained by performing the chromatographic analyses with Waters Alliance model 2690 HPLC with a refractive index detector Waters model 2410 (Waters, Milford, MA, USA). A Waters xbridge C18 column with 4.6 mm × 250 mm internal diameter with 5 μm particle size was used to achieve the chromatographic separation of the compounds in the sapucaia oil. Instrument settings were as follows:

2.2. Sapucaia nut oil extraction The oil samples utilized in this study were obtained by three different methods: subcritical propane extraction (LPP1 to LPP5), supercritical CO2 using ethanol (1:1, w/w) as co-solvent (LPC), and Soxhlet using n-hexane (LPS). The oil extracted by Soxhlet was mainly used as control. All the extraction conditions are summarized in Table 1. 123

The Journal of Supercritical Fluids 133 (2018) 122–132

G.L. Teixeira et al.

Fig. 1. Sampling of sapucaia nuts and the resulting extracted oils. A: whole sapucaia nuts; B: detail of a crushed nut; C: peeled nuts; D: crushed nuts separated by #8 mesh sieve; E: crushed nuts separated by #12 mesh sieve; F: residue of nut sample after oil extraction; G: oil appearance; H: oil samples stored in amber glass vessels.

5 °C/min then held for 5 min followed by cooling to −80 °C at −5 °C/ min and held at this temperature for 5 min. After that, samples were heated from −80 to 50 °C at 5 °C/min. Thermogram was registered, and peak and onset temperatures were calculated using STARe software (Mettler Toledo).

isocratic flow (1 mL/min), temperature of 40 °C for sample, column chamber and detector, and a mobile phase containing acetone/acetonitrile 6:4 (v/v) (mixed and de-gassed in advance). Millenium32 software (K & K Testing, LLC, Decatur, GA, USA) was used to analyze the data. TAG composition was determined using the internal library for each TAG and its corresponding signal.

2.6. Oxidative stability index 2.5. Thermal behavior of sapucaia oils The oxidative stability index (OSI), also known as Induction Period (IP), was obtained in a Metrohm Rancimat model 679 (Herisau, Switzerland), following the AOCS Official Method Cd 12b-92 [26]. In short, the deionized water conductivities were continually measured while 20 L/h of air was bubbled into 2.5 ± 0.1 g of oil sample, which was heated at 110 °C, and meanwhile the volatile compounds formed

TA Mettler-Toledo differential scanning calorimeter (DSC) (Mettler Toledo, Mississauga, ON, Canada) was used to determine thermal behavior. Oil samples ranging from 10 to 12 mg were weighed and hermetically sealed into an aluminum pan. A cooling and heating test was performed. Samples were held at 20 °C for 5 min and heated to 50 °C at

Table 1 Experimental conditions and results for extraction of sapucaia (Lecythis pisonis) nut oils with subcritical propane, supercritical CO2 with ethanol as co-solvent, and n-hexane. Run

Sample

Method

Solvent

T (K)

P (MPa)

Time of extraction (min)

Extraction yield

1 2 3 4 5 6

LPP1 LPP2 LPP3 LPP4 LPP5a LPC

Subcritical fluid extraction

293.15 333.15 293.15 333.15 313.15 333.15

2 2 10 10 6 20

60 60 60 60 60 60

42.49 43.61 42.35 46.22 43.47 ± 0.72 33.32 ± 0.50

7

LPS

Propane Propane Propane Propane Propane CO2 + ethanol (1:1 wt/wt %) n-hexane

Boiling point



360

49.50 ± 0.51

a b c d

Supercritical fluid extraction Soxhlet

b b

b

c

(wt%)

Extraction efficiency (%) d 85.85 88.10 85.56 93.38 87.81 67.32 100

center point. mean ± standard deviation (n = 3). mass of extract by the mass of dried material fed × 100. mass of extract obtained with the pressurized solvents at the end of the extraction time by mass of extract obtained with n-hexane in the classical extraction × 100.

124

The Journal of Supercritical Fluids 133 (2018) 122–132

G.L. Teixeira et al.

Fig. 2. Schematic diagram of the extraction unit used in this work. C1: Solvent cylinder; B1 and B2: circulation baths; P1: syringe pump; E1: extractor; S: sampling trap; V1: ball valve; V2 and V3: needle valves; V4: micro-metering needle valve; dashed lines: heat exchanger fluids.

The oil removal from sapucaia nut was almost completed during the first 40 min of the process, and during this period the extraction rate increased almost linearly over time due to the mass transfer being facilitated by the extraction on particles’ surface [28]. Extraction yield obtained with n-hexane (Soxhlet extraction) was higher than the other solvents, but it is noteworthy that in this case the extraction time was 360 min, while for the compressed fluid extractions the time was 6 × faster, verifying that a high oil yield can be obtained with short extraction times. Furthermore, subcritical propane achieved ∼93 wt% of the yield obtained by n-hexane, while the other scCO2 achieved ∼67.32 wt% (Table 1), suggesting the former as the best option for oil recovery from sapucaia nuts. Compressed propane has already been proved to be a better solvent for extracting oils than CO2 [9,11,22].

after decomposition of hydroperoxides were collected in water. The time taken to reach the conductivity inflection point was recorded. The IP was registered by Rancimat 679 printer as “hours”. 2.7. Statistical analysis Fatty acids, triacylglycerol and thermal analyses were performed in duplicate. Means and standard deviations were calculated and the results were submitted to Tukey’s and Duncan’s tests at 5% probability. Once the difference between the results was more accurately described by Duncan’s, this was chosen for the main statistical test. Analysis of variance was used to evaluate the effect of independent variables on the responses using mathematical model expressed by Eq. (1) (Ri = response, β0 = constant, and β1, β2 and β12 = regression terms; T refers to temperature, P refers to pressure). All statistical evaluation was done by using Statistica 10.0 (StatSoft Inc.). Graphs were made with the aid of Origin 8.6 (Originlab). Ri = β0 + β1T + β2P + β12TP

3.2. Statistical modeling Fig. 3a and b represents the temperature influence on the kinetics of extraction at two different fixed pressures, while Fig. 3c and d refers to the pressure effect on the extraction kinetics. Fig. 4a depicts the response surface obtained using the linear model (Eq. (1)), and Fig. 4b is the Pareto chart with the effect of the independent variables on the total ye, from ANOVA analysis. In Table 2 the parameters of the propane, including densities, viscosities and apparent solubility in each extraction condition are presented. As shown in Fig. 3, even the lowest pressure applied was enough to achieve a high initial rate and almost the entire oil content was extracted in a few minutes. From an industrial point of view, a short extraction time is of great interest, since it can also signify reduction of costs, faster production, and lower energy inputs. In addition, extraction at a low temperature is also important, because some temperature sensitive components can be preserved after the extraction period. As mentioned, all evaluated conditions showed high initial extraction rates, and the maximum extraction was reached at around 60 min. It is observed from Figs. 3 and 4 that the temperature had a more pronounced effect in the ye (Fig. 4b) at 10.0 MPa (Fig. 3b) than at 2.0 MPa (Fig. 3a), since under studied conditions, propane is a compressed liquid and changes in its density with variations in the pressure are small (see Table 2). Fig. 3a and b evidence that an increase in the temperature favors the initial rates of extraction. The Eq. (2) (R2 = 0.94) confirmed that only temperature (bold) contributed significantly (p < 0.05) to an increase in the ye of sapucaia oils. The ANOVA (Table 3) also indicates that the model was significant, proving that temperature may influence the selectivity of lipid fractions in sapucaia oil. This can be explained, because an increase in the temperature can cause a reduction in the viscosity of the solvent (Table 2), an increase in vapor pressure of the extracted components and an increase in diffusivity [5,13,22], thus diminishing the mass

(1)

3. Results and discussion 3.1. Extraction yield The fixed-bed for pressurized extractions was formed with 30.0 ± 0.3 g of sapucaia nut with a moisture content of 4.07 ± 0.90 wt%. The particle size (average particle diameter) was around 2.36 × 10−3 m. Table 1 shows the experimental conditions applied for the extraction of sapucaia nut oils (SNO) using compressed propane, scCO2 (with 1:1 wt/wt% ethanol to raw material ratio) and n-hexane as solvents. For the extraction yield (ye) calculation, the mass of extract was divided by the mass of dried material fed, after the extraction period (60 min), in order to allow a direct comparison between the obtained results for the different experimental conditions, and the results are reported as extraction percentage (wt%). In addition, Table 1 gives a comparison in the ye using the ye value obtained with n-hexane in the Soxhlet as a reference value, adopted as 100 wt% of extraction, to calculate the extraction efficiency for propane and scCO2. Using n-hexane as solvent resulted in an ye of 49.50 wt% of oil. The highest ye (46.22 wt%) using propane was obtained at higher levels of temperature (333.15 K) and pressure (10 MPa) in a short-period extraction (60 min). Thus, this is the most efficient experimental condition, providing a high initial rate and the highest final ye. The lowest yield (33.32 wt%) was obtained using scCO2 + ethanol (1:1 wt/wt%) as solvent and using a high pressure (20 MPa); a result ∼28% lower compared to the higher yield obtained by using propane (LPP4). 125

The Journal of Supercritical Fluids 133 (2018) 122–132

G.L. Teixeira et al.

Fig. 3. Influence of temperature (a, b) and pressure (c, d) on the extraction of sapucaia nut oils using subcritical propane (LPP1 to LPP5); a comparison of all the kinetic of subcritical propane extraction with the sample obtained by supercritical CO2 with 1:1 (w/w) ethanol as co-solvent (LPC) (e); and extraction yield comparison to Soxhlet using n-hexane (f).

Fig. 4. Response surface (a) and Pareto chart (b) comparing the effects of the temperature and pressure on the extraction yield of sapucaia nut oil obtained by using compressed propane.

126

The Journal of Supercritical Fluids 133 (2018) 122–132

G.L. Teixeira et al.

note that the conditions of subcritical propane used in the extractions resulted in values of apparent solubility (Sap) ranging between 0.596–0.734 gextract/gsolvent, evidencing that there is a high level of oil which is easily accessed by the solvent. The extraction conditions for sample LPP4 provided a high degree of solubility (Sap = 0.734 goil/ gsolvent), resulting in higher amounts of extracted oil. Experimental results demonstrated that 1.0 kg of propane can afford up to 734.0 g of sapucaia nut oil. Silva et al. [18] determined the solubility of perilla oil in compressed propane under the conditions of 40, 60 and 80 °C, and 8, 12 and 16 MPa. Values were 0.40–0.48 g oil/g solvent, which are close to those obtained in this study. The percentage of oil in sapucaia nut can vary due to many factors, including agronomic conditions and region of production. Vallilo et al. [1] have found almost 50% variation in the results of oils extracted from sapucaia nuts obtained from different regions in Brazil (34.2–61.3 wt% of oil). In the present study, the oil yields were close to the values reported by Vallilo et al. [1] and lower than those reported by Costa & Jorge [2] and Carvalho et al. [24]. In our previous study [25] using Bligh-Dyer and Soxhlet (with n-hexane) techniques, the oil yields were similar to this work (49.50 and 55.20 wt%, respectively. The compressed propane showed excellent solubility in sapucaia nut, and the extraction has proved to be very effective, presenting extraction yields slightly lower than those obtained by the conventional method (Fig. 3f). It can also be stated that the SPE extraction has satisfactory yields when compared to scCO2+ethanol, which the difference between the highest yield (LPP4) was ∼26%. Furthermore, it is noteworthy that the extraction time using subcritical propane is much less than for Soxhlet extraction. Besides, another advantage that the subcritical extraction presents is that the solvent is rapidly separated from the oil simply by releasing the pressure. On the contrary, Soxhlet and scCO2 + ethanol extractions need one more step to evaporate the solvent from the oil, which requires higher energetic efforts without guaranteeing a totally free-solvent product. The iodine values (IV) reflect the amount of unsaturation that the oil presents. Higher values of this parameter indicate higher degree of unsaturation, with higher probability of being a liquid oil at ambient temperature instead of a fat. In addition, the IV value could reflect the susceptibility of oil to oxidation. The IV found in the SNO samples obtained in this work (94.71–104.70 mg I2/100 g; see Table 4) agree with the fatty acid composition, which shows a high degree of unsaturated FA [1,2]. The results reported herein are higher than the values reported by Costa & Jorge [2], and similar to those reported by Vallilo et al. [1]. The IV for sapucaia nut oil are similar to the ones for Brazil nut oil (103 mg I2/100 g) [2] corn oil (103–128 mg I2/100 g), cottonseed oil (99–119 mg I2/100 g), and peanut oil (80–106 mg I2/ 100 g) [30].

Table 2 Parameters for propane in different conditions used in the extraction of oils from sapucaia nut. Sample

T (K)

P (MPa)

Density (g/ cm3)1

Viscosity (μPa s)1

Apparent solubility (goil/gsolvent)2

LPP1 LPP2 LPP3 LPP4 LPP5

293.15 333.15 293.15 333.15 313.15

2.0 2.0 10.0 10.0 6.0

0.503 0.428 0.521 0.467 0.484

104.610 65.682 118.760 83.293 92.417

0.596 0.684 0.666 0.734 0.629 ± 0.014

a

a = average ± standard deviation. 1 Data obtained from NIST Chemistry WebBook. 2 calculated as recommended by Sovová [29], by a linear plot of the initial extraction points (chart). Table 3 Analysis of variance (ANOVA) for the linear model of the overall extraction yield of sapucaia nut oil obtained by using subcritical propane. Source

Sum of squares

Degrees of freedom

Mean square

F value

p-value

T P T.P Lack-of-fit Pure error Total

6.219 1.524 1.899 0.071 0.524 10.237

1 1 1 1 2 6

6.219 1.524 1.899 0.071 0.262 –

23.740 5.817 7.249 0.270 – –

0.040* 0.137 0.115 0.655 – –

T, Temperature; P, Pressure. *Significant at 95% confidence level.

transfer resistance, enhancing the transport properties (as the diffusion coefficient) and facilitating the oil extraction. It is noteworthy that some lipids, such as oleic acid, present an increase in solubility with increasing temperature [28]. Moreover, in the subcritical region, instead of pressure, temperature is the determinant factor. The positive effect of temperature on subcritical extraction using propane of oils from sesame seeds [11], canola seeds [14], perilla [18], and sunflower seeds [23] has been previously reported. Although pressure had a positive effect in the ye, the results showed that changes in this parameter have no significant effect (Figs. 3 c, d, and 4 a, b). The same behavior has been reported by other authors for different solid matrixes [5,11]. Additionally, at 313.15 K and 6.0 MPa (center point), the kinetics of extraction showed results close to the other conditions (Fig. 3d). Ri = 39.3155 + 0.0107T − 2.5428P + 0.0086TP

(2)

The kinetic extraction curves of sub- and supercritical extractions present similar behavior at the beginning of extraction, possibly because the solvent is saturated with the extract. Analyzing all the curves together (Fig. 3e), after some points, it is visible that as the temperature rises they do not overlay, and some of them are higher than the others, which shows that the oil removal becomes easier (Fig. 3). The extraction yield reaches a plateau after a short period, indicating a high solubility of sapucaia oil in the solvent, which has been reported as a common behavior when using compressed propane to extract vegetable oils [4,19,22]. Mostly, the fluctuations in the initial rate of extraction were because of the variation in the density of propane (428.05–521.29 kg/m3), which is dependent on the temperature and pressure (Table 2). Moreover, the final yield extraction of oils can also be affected by density of the solvent. As a result, changes in these parameters can directly affect the kinetics of the extraction [3,5,6,8]. As explained by Sovová [29], in the static extraction, which occurs before opening of the extractor’s valve in the outlet at t = 0, the equilibrium was established, and from the initial slope of cumulative extraction curve the apparent solubility can be obtained. The oil solubility in the solvent was determined using the dynamic method, and was calculated from the slope of the linear part of the curves of overall extraction [29](Fig S1). From Table 2 it is possible

3.2.1. Oxidative stability Through Rancimat analysis at 110 °C, the Induction Period (IP) of Table 4 Iodine value and oxidative stability parameters of sapucaia nut oils obtained by subcritical propane (LPP1-LPP5), supercritical CO2 with 1:1 (w/w) ethanol as co-solvent (LPC), and n-hexane (LPS). Sample LPP1 LPP2 LPP3 LPP4 LPP5 LPC LPS

IV (mg I2/100 g)1 a

104.70 ± 0.03 104.57 ± 0.04a 104.62 ± 0.14a 104.69 ± 0.05a 104.59 ± 0.31a 104.29 ± 0.01a 94.71 ± 0.13b

IP (h)

PF a

10.80 ± 0.71 11.17 ± 2.31a 7.38 ± 0.29bc 7.34 ± 0.62bc 6.53 ± 0.43c 9.29 ± 0.59ab 7.44 ± 0.06bc

1.541 1.500 0.991 0.986 0.877 1.248 –

± ± ± ± ± ±

0.084a 0.299a 0.047bc 0.075bc 0.065c 0.088ab

1 Calculated by difference from fatty acids profile according to AOCS Cd 1c-85; IV, Iodine value; IP, Induction Period; PF, Protection Factor (dimensionless). Results are the mean ± standard error (n = 2); means followed by same letter do not differ by Duncan test (p < 0.05).

127

The Journal of Supercritical Fluids 133 (2018) 122–132

G.L. Teixeira et al.

levels of linoleic acid, a decrease of approximately 28% in comparison to the other solvents. On the other hand, n-hexane provides an oil with ∼18% higher quantity of oleic acid. This could be related to the polarity of the solvent, which has more affinity with certain FA. For example, oleic acid, which is a long-chain FA with one double bond, has lower polarity compared to other fatty acids. For this reason, the extraction using n-hexane provided higher yield for this component. For medium-chain FA (palmitic acid, for example), solvents with high polarity seem to afford higher extraction efficiency. For example, for lipid extraction from yeasts, the type of solvent, alcohol, ester, terpene or ether, has a significant impact in the FA composition [35]. The same hold true for hexane, scCO2, chloroform and methanol, which present different behavior in the extraction of lipids from microalgae [36]. The FA composition of the sapucaia nut oil obtained in the study reported herein is similar to others reported in the literature, such as Vallilo et al. [1], Costa & Jorge [2] and Carvalho et al. [24]. In these studies, the levels of oleic and linoleic acid varied from 38.82 to 43.0% and from 39.90 to 56.30%, respectively. It is noteworthy that the FA composition of SNO is quite comparable to that of Brazil nut oil [31]. Representing between 20.75–22.11% of the total FA, the saturated fatty acids (SFA) from sapucaia oils were mainly palmitic acid (13.95–15.22%) and stearic acid (5.45–8.00%). Myristic and margaric acids were identified in smaller amounts, while behenic and lignoceric acids were traces. Different FA compositions were also found for sesame seed oil according to different solvents used in the Soxhlet system. Higher amounts of stearic and oleic acids were found when using petroleum ether; using ethanol, palmitic and linoleic were the major different values [37]. As a matter of fact, by using propane at different conditions of T and P, the SNO showed the same composition of FA. Within the parameters of this study, therefore, propane does not provide changes in the distribution of FA, also proving that it does not promote any lipid transformation, such as occurs in rice brain oil [28]. The advantages of intake of nut oils such as salad oil are also associated with reduced risk of coronary heart disease, diabetes, and cancer [38].

sapucaia oils was measured, and the Protection Factor (PF) was then calculated from IP results, both presented in Table 4. The SNO presents significant differences (p < 0.05) among the values of IP and PF according to the extraction method. Oil samples extracted using n-hexane showed ∼7.44 h of stability, while those extracted using propane presented values between 6.53–11.17 h. In addition, oil samples extracted using scCO2+ethanol showed good stability against oxidation, with approximately 9.29 h of IP. Oils extracted using subcritical propane at low pressure (2 MPa) were the samples that presented better stability, showing the highest induction periods. On the contrary, samples obtained at higher pressure (6 or 10 MPa) presented the lowest IP values, indicating poor oxidative stability. The values for PF, which are related to the antioxidant capacity of the oils, also confirmed that samples extracted using compressed solvents presented good antioxidant properties (PF > 1.0), and the ones which were extracted at higher pressure values were less effective against oxidation (PF < 1.0). Teixeira et al. [25] showed similar IP results for SNO extracted by Soxhlet (7.18 h) and Bligh & Dyer (13.00 h) techniques. An IP of 24.89 h was reported for SNO extracted using cold pressing, but the evaluation in Rancimat in this case was conducted at 100 °C [2]. Our results are higher than the IP for sunflower oil (4.91 h), and close to macadamia (7.38 h), Brazil nut (8.24 h), canola (8.63 h), hazelnut (8.88 h), pecan (9.87), corn (9.96 h), and soybean (12.0 h) oils [31].

3.2.2. Fatty acid composition of sapucaia oils The fatty acid (FA) compositions of the obtained oils are shown in Table 5. In this study, the FA profiles were practically the same despite the different extraction methods used, except for the sample extracted using n-hexane. Eleven different FA were identified. Unsaturated fatty acids (UFA) were predominant (77–79%) in all samples, particularly similar to Brazil nut [32–34]. Monounsaturated fatty acids (MUFA) accounted for 42.06–50.85%; polyunsaturated fatty acids (PUFA), mainly linoleic (omega-6) and α-linolenic acids, were between 26.77–36.60%. Regarding the MUFAs, the major FA identified in these sapucaia oil samples was oleic acid, which accounted for ∼50% in LPS and ∼42% in the other samples. Palmitoleic and gondoic acids were identified in amounts less than 0.2%. The statistical analysis showed a significant difference (p > 0.05) between the different solvents used in relation to the FA composition, particularly the amount of oleic, linoleic and palmitic acids, which are the major FA in these oils. Extraction using n-hexane results in lower

3.2.3. Thermal behavior of sapucaia oils The crystallization and melting curves of the studied sapucaia oils are shown in Fig. 5, and the parameters for each thermal event are shown in Tables 6 and 7 (crystallization and melting, respectively). During cooling of SNO, two exothermic peaks were detected (Table 6,

Table 5 Fatty acid profile of sapucaia nut oils obtained by subcritical propane (LPP1-LPP5), supercritical CO2 with 1:1 (w/w) ethanol as co-solvent (LPC), and n-hexane (LPS). Fatty acids

Myristic (C14:0) Palmitic (C16:0) Palmitoleic (C16:1) Margaric (C17:0) Stearic (C18:0) Oleic (C18:1) Linoleic (C18:2) α-Linolenic (C18:3 Gondoic (C20:1) Behenic (C22:0) Lignoceric (C24:0) Others Σ SFA 1 Σ MUFA 2 Σ PUFA 3 Σ PUFA + MUFA PUFA/SFA 1 2 3

Composition (%) LPP1

LPP2

LPP3

LPP4

LPP5

LPC

LPS

0.09 ± 0.01a 15.22 ± 0.00ab 0.22 ± 0.01bc 0.06 ± 0.01a 5.53 ± 0.01c 42.18 ± 0.04cd 35.83 ± 0.02abc 0.61 ± 0.01a 0.09 ± 0.00ab tr tr 0.11 ± 0.01a 20.90 ± 0.01bc 42.49 ± 0.03d 36.43 ± 0.02abc 78.93 ± 0.01a 1.75

0.08 ± 0.01a 15.06 ± 0.13b 0.22 ± 0.01bc tr 5.58 ± 0.02b 42.71 ± 0.19b 35.52 ± 0.09d 0.60 ± 0.02a 0.07 ± 0.00b tr tr 0.14 ± 0.02a 20.75 ± 0.06c 43.00 ± 0.19b 36.11 ± 0.11d 79.11 ± 0.08a 1.74

0.09 ± 0.00a 15.17 ± 0.03ab 0.22 ± 0.01bc 0.07 ± 0.00a 5.47 ± 0.01d 42.20 ± 0.05c 35.79 ± 0.07abc 0.60 ± 0.01a 0.09 ± 0.01ab tr tr 0.20 ± 0.00a 20.86 ± 0.03bc 42.50 ± 0.05d 36.39 ± 0.07abc 78.89 ± 0.03a 1.74

0.09 ± 0.01a 15.15 ± 0.04ab 0.27 ± 0.04ab 0.07 ± 0.00a 5.45 ± 0.01d 42.01 ± 0.10cd 35.89 ± 0.05ab 0.60 ± 0.00a 0.09 ± 0.00ab tr 0.05 ± 0.02 0.28 ± 0.01a 20.83 ± 0.05bc 42.36 ± 0.07cd 36.49 ± 0.05ab 78.85 ± 0.02a 1.74

0.12 ± 0.04a 15.16 ± 0.06ab 0.21 ± 0.01c tr 5.45 ± 0.02d 41.73 ± 0.16d 36.00 ± 0.02a 0.59 ± 0.02a 0.13 ± 0.04a tr tr 0.57 ± 0.24a 20.76 ± 0.07c 42.06 ± 0.13c 36.60 ± 0.05a 78.66 ± 0.17a 1.76

0.09 ± 0.01a 15.31 ± 0.01a 0.28 ± 0.01a 0.06 ± 0.01a 5.47 ± 0.00d 42.15 ± 0.01cd 35.61 ± 0.00cd 0.60 ± 0.01a 0.08 ± 0.00ab tr tr 0.22 ± 0.01a 20.97 ± 0.01b 42.50 ± 0.00d 36.21 ± 0.00cd 78.71 ± 0.00a 1.73

0.07 ± 0.00a 13.95 ± 0.02c 0.23 ± 0.00abc 0.08 ± 0.01a 8.00 ± 0.02a 50.53 ± 0.12a 26.17 ± 0.01e 0.59 ± 0.00a 0.09 ± 0.00ab tr tr 0.27 ± 0.13a 22.11 ± 0.02a 50.85 ± 0.12a 26.77 ± 0.01e 77.62 ± 0.12b 1.21

Saturated fatty acids. Monounsaturated fatty acids. Polyunsaturated fatty acids; Results are the mean ± standard deviation (n = 2); tr: percentage < 0.05; means followed by same letter do not differ by Duncan test (p < 0.05).

128

129

1.77b 5.19ab 5.68a 9.33b 1.45ab 3.44ab 1.41ab ± ± ± ± ± ± ± 9.59 7.36 6.94 7.16 7.41 7.95 5.42 −64.66 −63.66 −62.11 −63.12 −65.87 −63.51 −58.38 −57.50 −56.75 −56.09 −56.57 −57.55 −56.27 −53.14 −50.87 −50.89 −50.36 −50.76 −51.54 −49.76 −49.17 2.12 1.19 0.76 0.89 1.28 0.92 2.77 −12.28 ± 0.03d −10.29 ± 0.09c −7.84 ± 0.04ab −7.86 ± 0.02ab −10.42 ± 0.16c −7.67 ± 0.16a −8.44 ± 0.44b −9.92 ± 0.00e −8.84 ± 0.09d −7.00 ± 0.05c −7.04 ± 0.16c −8.91 ± 0.09d −6.58 ± 0.12b −4.70 ± 0.00a −8.65 ± 0.00e −8.22 ± 0.06d −6.59 ± 0.03c −6.48 ± 0.11c −8.20 ± 0.03d −6.14 ± 0.06b −4.49 ± 0.02a LPP1 LPP2 LPP3 LPP4 LPP5 LPC LPS

PCM TConset

TCendset

Pwidth

± ± ± ± ± ± ±

0.01b 0.02cd 0.07e 0.02de 0.08c 0.07de 0.24a

TConset

Peak 2

± ± ± ± ± ± ±

0.07d 0.08d 0.01c 0.00cd 0.08e 0.34b 0.09a

PCM

± ± ± ± ± ± ±

0.04d 0.11c 0.08b 0.05bc 0.14d 0.42bc 0.04a

TCendset

± ± ± ± ± ± ±

2.42b 0.06b 0.00ab 0.02b 1.48b 0.93b 0.08a

PCwidth

0.19a 0.08b 0.15b 0.01b 0.04b 0.71b 0.04c

−8.09 −7.68 −9.94 −5.38 −8.84 −7.70 −7.36

± ± ± ± ± ± ±

0.07ab 1.05ab 1.45b 2.37a 0.15ab 0.36ab 0.08ab

HTab (J/g) at 5.5 °C

18.73 22.60 25.62 19.09 23.34 21.36 25.99

± ± ± ± ± ± ±

HPM (J/g)

0.19b 1.49ab 1.22a 2.97b 0.38ab 0.80ab 0.07a

67.68 74.80 85.69 64.24 78.47 73.52 78.20

± ± ± ± ± ± ±

HTotal (J/g)

Fig. 5a). A shallow peak was obtained at around −4.49 to −8.65 °C (Ton). At this point, the crystallization of TAGs should occur, once they are crystallized at higher temperatures, and this first peak can be attributed to an oil fraction that contains mainly saturated fatty acids such as palmitic and stearic acid [39]. On the other hand, the unsaturated fatty acids which account for ∼79% of the oil are fully crystallized at the second deep peak (−49.17 to −51.54 °C) (Tendset). The absolute enthalpy (HTab, −5.38 to −9.94 J/g), the enthalpy for the main peak (HPM, 18.73–25.99 J/g), and the total enthalpy (HTotal, 64.24–85.69 J/g) presented significant differences (p < 0.05) between the samples, showing that the extraction methods could be responsible for changes in the thermal behavior of the sapucaia nut oils. Melting curves of sapucaia nut oils (Fig. 5b) consisted of a major peak during the melting process, accompanied by a major shoulder. Results showed that the melting of SNO started from −19.35 to −13.75 °C. Main peak was observed in the temperature range of −6.21 to −7.89 °C. Endset temperatures for SNO were found to be from 0.40 to −3.69 °C, where the oils are supposed to be fully liquid. The process comprised a melting absolute enthalpy range of 118.41–165.53 J/g, a main peak enthalpy between −2.35 to −18.37 J/g, and a total enthalpy of −59.00 to 74.97 J/g. Results indicated that the melting behavior of sapucaia nut oil could be significantly dependent upon the extraction method and the extraction temperature (293 K to 333 K). It is noticed that the sample extracted using n-hexane presented the most differences in thermal behavior, being the only sample to be fully liquid

Peak 1

Table 6 DSC parameters for crystallization (40 to −80 °C) behavior of sapucaia oils obtained by subcritical propane (LPP1-LPP5), supercritical CO2 with 1:1 (w/w) ethanol as co-solvent (LPC), and n-hexane (LPS).

Fig. 5. DSC curves showing (a) crystallization (40 to −80 °C) and (b) melting (−80 to 40 °C) behaviors of sapucaia oils obtained using subcritical propane (LPP1-LPP5), supercritical CO2 with 1:1 (w/w) ethanol as co-solvent (LPC), and n-hexane (LPS).

Results are the mean ± standard deviation (n = 2); means followed by same letter do not differ by Duncan test (p < 0.05). TConset, onset temperature for crystallization; PCM, crystallization temperature for the peak; TCendset, endset temperature for crystallization; Pwidth, width of the peak; HTab, total absolute enthalpy; HPM, enthalpy of the main peak; Htotal, total enthalpy.

The Journal of Supercritical Fluids 133 (2018) 122–132

G.L. Teixeira et al.

The Journal of Supercritical Fluids 133 (2018) 122–132

G.L. Teixeira et al.

Table 7 DSC parameters for melting (-80 to 40 °C) behavior of sapucaia oils obtained by subcritical propane (LPP1-LPP5), supercritical CO2 with 1:1 (w/w) ethanol as co-solvent (LPC), and nhexane (LPS). Sample

TMonset

LPP1 LPP2 LPP3 LPP4 LPP5 LPC LPS

−14.74 −14.73 −14.10 −14.36 −15.87 −13.75 −19.35

PMM (°C) ± ± ± ± ± ± ±

0.27b 0.04b 0.39ab 0.04ab 0.15c 0.42a 0.22d

−6.21 −7.59 −7.89 −7.60 −7.50 −7.42 −7.80

± ± ± ± ± ± ±

0.07a 0.01bcd 0.00d 0.10bcd 0.04bc 0.22b 0.04cd

Tendset (°C)

Pwidht (°C)

−3.69 ± 0.18d −2.46 ± 0.06b −2.90 ± 0.09c −2.81 ± 0.01c −2.37 ± 0.08b −2.67 ± 0.12bc 0.40 ± 0.01a

16.19 16.09 15.81 15.33 16.41 15.11 14.21

± ± ± ± ± ± ±

HTab (J/g) at 20 °C 0.19a 0.02ab 0.05b 0.14c 0.04a 0.05c 0.02d

151.05 143.26 174.67 118.41 165.53 149.28 148.88

± ± ± ± ± ± ±

2.06ab 19.52ab 9.79a 29.38b 4.30ab 6.31ab 1.75ab

HPM (J/g)

HTotal (J/g)

−2.35 ± 0.38a −6.56 ± 0.18b -9.45 ± 0.18c −7.13 ± 1.05b −6.39 ± 0.86b −7.53 ± 0.39b −18.37 ± 0.05d

−59.00 −68.97 −73.41 −61.35 −70.40 −68.64 −74.97

± ± ± ± ± ± ±

4.16a 3.02abc 1.75c 5.65ab 0.83bc 1.60abc 0.60c

Results are the mean ± standard deviation (n = 2); means followed by same letter do not differ by Duncan test (p < 0.05). TMonset, onset temperature for melting; PMM, melting temperature for the main peak; TCendset, endset temperature for melting; Pwidth, width of the peak; HTab, total absolute enthalpy; HPM, enthalpy of the main peak; Htotal, total enthalpy.

(13.71–14.91%) SOO (11.79–15.10%), POO (9.80–12.77%), PLO (8.50–12.51%) and POS (11.19–12.95). The major FA constituting TAG were oleic, linoleic, palmitic and stearic acids. This result is in accordance with GC-analysis of the total FA content of SNO, in which the composition of oleic acid (42–50%), linoleic acid (26–36%), palmitic acid (14–15%) and stearic acid (5–8%) are predominant (Table 5). This data reinforces the idea that fatty acids such as oleic (C18:1) or linoleic (C18:2) in vegetable oils are exclusively at sn-2 position in TAG species [43]. Moreover, the TAG composition of sapucaia nut oil reveals that this is another similarity to Brazil nut [32], and more surprisingly to olive oil, the main TAGs of which are OOO (21.8–43.1%) and POO (20–23.1%) [42], and also to rice bran oil (major TAGs are PLO, PLL and OOO) [38]. Among the reasons for the variations in TAG composition in samples of the same species could be the collection of material in different geographic locations or during different growing seasons. In addition, oils with the same fatty acid composition may not have the same physical, chemical or physiological properties, because of the differences in the TAG composition, which also influences the digestion [38], absorption and transport in the human organism [43].

at a temperature above 0 °C. Changes in melting point suggest the formation of new TAG molecular species, including possibly creation of disaturated and trisaturated TAGs [40]. Papaya seed oil had thermal behavior comparable to that of sapucaia nut oils, both presenting similar a pattern of peaks during melting (one major peak at −3.5 to −1.8 °C) and crystallization (two peaks at around −10 °C and −45 to −59 °C) [41]. These results can also be compared to extra virgin olive oil [39], which present analogous behavior, with thermal parameters similar to the ones presented by the studied sample. Fig. 5 also reveals that although the sapucaia oil samples present similar behavior for both crystallization and melting, some major differences can be found in the sample extracted using n-hexane (LPS).

3.2.4. Triacylglycerol composition of sapucaia oils Triacylglycerols (TAGs) are considered the most abundant single lipid class [42], once it comprises more than 95% of fats and oils composition [38]. Some functional properties of oils do not depend only on their FA composition, but also on the distribution of these FA in the three positions of the glycerol backbone [42], which are responsible for many health and physiological implications as a consequence [38]. Also, the commercial value of a fat or oil is highly influenced by its TAG composition [42]. In addition, in many fat-based foods, some properties such as melting and crystallization behavior are directly impacted by the TAG composition of that fat/oil [43]. TAG structure has also a main role in some properties such as texture, plasticity and mouth feel. The suitable end use of a product is most often defined according to the TAG species in the fat/oil used [42]. Fig. 6 shows the typical chromatogram of TAGs for the sapucaia oils. Table 8 reveals that the TAGs present in SNO are predominantly OOO

4. Conclusions Sapucaia nut oil was obtained using subcritical propane, supercritical CO2 (with 1:1 wt/wt ethanol to raw material ratio) and conventional (Soxhlet) extraction techniques. When using SPE, the best results were obtained by applying high pressure and high temperature (10 MPa and 333 K), with a yield of 46.22 wt% being obtained under these conditions. The extraction of SNO with subcritical propane provides satisfactory extraction yields within 60 min, which represents Fig. 6. Typical HPLC chromatograms for TAG composition of sapucaia oils obtained by subcritical propane (a), supercritical CO2 with 1:1 (wt/wt%) ethanol as co-solvent (b), and n-hexane (c).

130

The Journal of Supercritical Fluids 133 (2018) 122–132

G.L. Teixeira et al.

Table 8 TAG composition of sapucaia nut oils obtained by subcritical propane (LPP1-LPP5), supercritical CO2 with 1:1 (w/w) ethanol as co-solvent (LPC), and n-hexane (LPS). TAG species

OLL PLL PLO PLP OOO POO POP SOO SLS POS PPS SOS PSS

Composition (%) LPP1

LPP2

LPP3

LPP4

LPP5

LPC

LPS

8.25 ± 0.17a 0.28 ± 0.01a 12.22 ± 0.06a 9.81 ± 0.04a 14.73 ± 0.10a 12.63 ± 0.08a 3.26 ± 0.14a 12.73 ± 0.59b 2.40 ± 0.66a 11.42 ± 0.01b 2.71 ± 0.13b 4.39 ± 0.23b 1.84 ± 0.01b

8.11 ± 0.06a 0.24 ± 0.01a 12.30 ± 0.16a 9.75 ± 0.13a 14.77 ± 0.13a 12.47 ± 0.03a 3.04 ± 0.03a 12.03 ± 1.30b 2.95 ± 0.98a 11.59 ± 0.47b 2.68 ± 0.23b 4.61 ± 0.01b 1.60 ± 0.63b

8.21 ± 0.06a 0.27 ± 0.07a 12.47 ± 0.06a 10.14 ± 0.04b 14.62 ± 0.05a 12.77 ± 0.13a 3.20 ± 0.18a 12.09 ± 0.85b 2.72 ± 0.66a 11.71 ± 0.04b 2.70 ± 0.04b 4.32 ± 0.13b 1.79 ± 0.11b

8.34 ± 0.13a 0.29 ± 0.01a 12.39 ± 0.29a 9.96 ± 0.12ab 14.91 ± 0.39a 12.68 ± 0.29a 3.10 ± 0.06a 12.23 ± 0.07b 2.53 ± 0.05a 11.19 ± 0.09b 2.59 ± 0.08b 4.58 ± 0.33b 1.92 ± 0.13b

8.19 ± 0.05a 0.30 ± 0.02a 12.51 ± 0.08a 9.96 ± 0.11ab 14.48 ± 0.11a 12.53 ± 0.09a 3.09 ± 0.06a 11.79 ± 0.00b 3.09 ± 0.18a 11.32 ± 0.06b 2.67 ± 0.14b 4.78 ± 0.54b 1.91 ± 0.02b

8.12 ± 0.05a 0.27 ± 0.01a 12.39 ± 0.00a 9.92 ± 0.12ab 14.73 ± 0.06a 12.72 ± 0.08a 3.05 ± 0.06a 12.37 ± 0.50b 2.49 ± 0.41a 11.44 ± 0.10b 2.72 ± 0.13b 4.57 ± 0.05b 1.90 ± 0.02b

4.15 ± 0.15b 0.12 ± 0.13b 8.50 ± 0.19b 5.15 ± 0.15c 13.71 ± 0.28b 9.80 ± 0.04b 2.02 ± 0.01b 15.10 ± 0.06a 2.86 ± 0.18a 12.95 ± 0.45a 3.18 ± 0.27a 7.12 ± 0.18a 3.35 ± 0.11b

O, Oleic acid; L, Linoleic acid; P, Palmitic acid; S, Stearic acid. Results are the mean ± standard deviation (n = 2); means followed by same letter do not differ by Duncan test (p < 0.05).

∼93% of the extraction yield obtained applying conventional extraction. Oleic and linoleic acids were predominant in the fatty acid composition of the oils obtained applying the different methods and conditions. Stability to oxidation measured by Rancimat ranged from 6.53 to 11.17 h. The methods of extraction using compressed solvents impacted as minor changes in the melting and crystallization behavior; however, n-hexane extraction caused more pronounced effects in these parameters. The triacylglycerol composition for sapucaia nut oil is presented here for the first time, and the TAGs are predominantly OOO, SOO, POO, PLO and POS. From the results obtained, it could be pointed out that propane is a more suitable solvent for sapucaia nut oil extraction than carbon dioxide, as higher extraction yields were achieved with this solvent. In general, this study highlights the application of compressed propane in an efficient extraction process of a good composition of oil (sapucaia oil), which has great potential for use in the food, pharmaceutical and chemical industries.

[6]

[7]

[8]

[9]

[10]

[11]

Acknowledgements [12]

This work was supported by the Coordination for the Improvement of Higher Education Personnel – CAPES (Brazil), grant n. 1291783 (CAPES-DS) and n. 88881.135997/2016-01 (PDSE). Thanks are also due to Prof. Dr. A. G. Marangoni for the supervision of G. L. Teixeira during the doctorate sandwich-period at the Food, Health and Aging Laboratory of the Department of Food Science at the University of Guelph (Canada).

[13]

[14]

Appendix A. Supplementary data [15]

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.supflu.2017.10.003.

[16]

References [17] [1] M.I. Vallilo, M. Tavares, S. Aued-Pimentel, N.C. Campos, J.M.M. Neto, Lecythis pisonis Camb.: oil characterization, fatty acids and minerals, Food Chem. 66 (1999) 197–200, http://dx.doi.org/10.1016/S0308-8146(99)00040-0. [2] T. Costa, N. Jorge, Characterization and fatty acids profile of the oils from Amazon nuts and walnuts: characterization and fatty acids profile of the oilseeds, Nutr. Food Sci. 42 (2012) 279–287, http://dx.doi.org/10.1108/00346651211248647. [3] K. Czaikoski, M.C. Mesomo, A. de Paula Scheer, O.R. Dalla Santa, C.L. Queiroga, M.L. Corazza, Kinetics, composition and biological activity of Eupatorium intermedium flower extracts obtained from scCO2 and compressed propane, J. Supercrit. Fluids 97 (2015) 145–153, http://dx.doi.org/10.1016/j.supflu.2014.10.008. [4] C.P. Trentini, K.A. Santos, E. Antonio da Silva, V.A. dos S. Garcia, L. Cardozo-Filho, C. da Silva, Oil extraction from macauba pulp using compressed propane, J. Supercrit. Fluids 126 (2017) 72–78, http://dx.doi.org/10.1016/j.supflu.2017.02. 018. [5] M. Correa, M.C. Mesomo, K.E. Pianoski, Y.R. Torres, M.L. Corazza, Extraction of

[18]

[19]

[20]

131

inflorescences of Musa paradisiaca L. using supercritical CO2 and compressed propane, J. Supercrit. Fluids 113 (2016) 128–135, http://dx.doi.org/10.1016/j. supflu.2016.03.016. K. Czaikoski, M.C. Mesomo, R.L. Krüger, C.L. Queiroga, M.L. Corazza, Extraction of Campomanesia xanthocarpa fruit using supercritical CO2 and bioactivity assessments, J. Supercrit. Fluids 98 (2015) 79–85, http://dx.doi.org/10.1016/j.supflu. 2015.01.006. M.G. Pereira, F. Hamerski, E.F. Andrade, A. de P. Scheer, M.L. Corazza, Assessment of subcritical propane, ultrasound-assisted and Soxhlet extraction of oil from sweet passion fruit (Passiflora alata Curtis) seeds, J. Supercrit. Fluids. (2017), http://dx. doi.org/10.1016/j.supflu.2017.03.021. R. Coelho, L.R.S. Kanda, F. Hamerski, M.L. Masson, M.L. Corazza, Extraction of kiwifruit seed (Actinidia deliciosa) oil using compressed propane, J. Food Process Eng. 39 (2016) 335–344, http://dx.doi.org/10.1111/jfpe.12225. P.M. Ndiaye, M. Lanza, F.W. Tavares, C. Dariva, D. Oliveira, J.V. Oliveira, Phase behavior of olive and soybean oils in compressed propane and N-butane, Braz. J. Chem. Eng. 23 (2006) 405–415, http://dx.doi.org/10.1590/S010466322006000300014. R.P.F.F. da Silva, T.A.P. Rocha-Santos, A.C. Duarte, Supercritical fluid extraction of bioactive compounds, TrAC Trends Anal. Chem. 76 (2016) 40–51, http://dx.doi. org/10.1016/j.trac.2015.11.013. M.P. Corso, M.R. Fagundes-Klen, E.A. Silva, L. Cardozo Filho, J.N. Santos, L.S. Freitas, C. Dariva, Extraction of sesame seed (Sesamun indicum L.) oil using compressed propane and supercritical carbon dioxide, J. Supercrit. Fluids 52 (2010) 56–61, http://dx.doi.org/10.1016/j.supflu.2009.11.012. A.A. Jesus, L.C. Almeida, E.A. Silva, L.C. Filho, S.M.S. Egues, E. Franceschi, M. Fortuny, A.F. Santos, J. Araujo, Sousa E.M.B.D, C. Dariva, Extraction of palm oil using propane, ethanol and its mixtures as compressed solvent, J. Supercrit. Fluids 81 (2013) 245–253, http://dx.doi.org/10.1016/j.supflu.2013.06.011. A.S. Pessoa, R. Podestá, J.M. Block, E. Franceschi, C. Dariva, M. Lanza, Extraction of pequi (Caryocar coriaceum) pulp oil using subcritical propane: determination of process yield and fatty acid profile, J. Supercrit. Fluids 101 (2015) 95–103, http:// dx.doi.org/10.1016/j.supflu.2015.03.006. M.M. Pederssetti, F. Palú, E.A. da Silva, J.H. Rohling, L. Cardozo-Filho, C. Dariva, Extraction of canola seed (Brassica napus) oil using compressed propane and supercritical carbon dioxide, J. Food Eng. 102 (2011) 189–196, http://dx.doi.org/10. 1016/j.jfoodeng.2010.08.018. B.M.V. Illés, H.G. Daood, P.A. Biacs, M.H. Gnayfeed, Supercritical CO 2 and subcritical propane extraction of spice red pepper oil with special regard to carotenoid and tocopherol content, J. Chromatogr. Sci. 37 (1999) 345–352. K.A. Santos, R.A. Bariccatti, L. Cardozo-Filho, R. Schneider, F. Palú, C. da Silva, E.A. da Silva, Extraction of crambe seed oil using subcritical propane: kinetics, characterization and modeling, J. Supercrit. Fluids 104 (2015) 54–61, http://dx. doi.org/10.1016/j.supflu.2015.05.026. A.B. Zanqui, D.R. De Morais, C.M. Da Silva, J.M. Santos, S.T.M. Gomes, J.V. Visentainer, M.N. Eberlin, L. Cardozo-Filho, M. Matsushita, Subcritical extraction of flaxseed oil with n-propane: composition and purity, Food Chem. 188 (2015) 452–458, http://dx.doi.org/10.1016/j.foodchem.2015.05.033. C.M. Da Silva, A.B. Zanqui, A.K. Gohara, A.H.P. De Souza, L. Cardozo-Filho, J.V. Visentainer, L.U. Rovigatti Chiavelli, P.R.S. Bittencourt, E.A. Da Silva, M. Matsushita, Compressed n-propane extraction of lipids and bioactive compounds from Perilla (Perilla frutescens), J. Supercrit. Fluids 102 (2015) 1–8, http://dx.doi. org/10.1016/j.supflu.2015.03.016. M.O. Silva, F.P. Camacho, L. Ferreira-Pinto, W.M. Giufrida, A.M.S. Vieira, J.V. Visentaine, D.R.L. Vedoy, L. Cardozo-Filho, Extraction and phase behaviour of Moringa oleifera seed oil using compressed propane, Can. J. Chem. Eng. 94 (2016) 2195–2201, http://dx.doi.org/10.1002/cjce.22614. A.B. Zanqui, C.M. da Silva, D.R. de Morais, J.M. Santos, S.A.O. Ribeiro, M.N. Eberlin, L. Cardozo-Filho, J.V. Visentainer, S.T.M. Gomes, M. Matsushita, Sacha inchi (Plukenetia volubilis L.) oil composition varies with changes in

The Journal of Supercritical Fluids 133 (2018) 122–132

G.L. Teixeira et al.

[21]

[22]

[23]

[24]

[25]

[26] [27] [28]

[29]

[30]

[31]

[32]

(2012) 11263–11267. [33] V. Queiroga Neto, O.A. Bakke, C.M.P. Ramos, P.S. Bora, J.C. Letelier, M.M. Conceição, Brazil nut (Bertholletia excelsa HBK) seed kernel oil: characterization and thermal stability, Rev. Bras. Biol. E Farmácia 3 (2009) 33–42 http:// sites.uepb.edu.br/biofar/download/v3n2-2009/04-BRAZIL_NUT.pdf. [34] O.V. Santos, N.C.F. Corrêa, F.a.S.M. Soares, L.a. Gioielli, C.E.F. Costa, S.C.S. Lannes, Chemical evaluation and thermal behavior of Brazil nut oil obtained by different extraction processes, Food Res. Int. 47 (2012) 253–258, http://dx.doi.org/10.1016/ j.foodres.2011.06.038. [35] C. Breil, A. Meullemiestre, M. Vian, F. Chemat, Bio-based solvents for green extraction of lipids from oleaginous yeast biomass for sustainable aviation biofuel, Molecules 21 (2016) 1–14, http://dx.doi.org/10.3390/molecules21020196. [36] Y. Li, F. Ghasemi Naghdi, S. Garg, T. Adarme-Vega, K.J. Thurecht, W. Ghafor, S. Tannock, P.M. Schenk, A comparative study: the impact of different lipid extraction methods on current microalgal lipid research, Microb. Cell Fact. 13 (2014) 14, http://dx.doi.org/10.1186/1475-2859-13-14. [37] R.H.R. Carvalho, E.L. Galvão, J.A.C. Barros, M.M. Conceição, E.M.B.D. Sousa, Extraction, fatty acid profile and antioxidant activity of sesame extract: (Sesamum Indicum L.), Braz. J. Chem. Eng. 29 (2012) 409–420, http://dx.doi.org/10.1590/ S0104-66322012000200020. [38] S.M. Ghazani, A.G. Marangoni, Healthy Fats and Oils, 2nd ed., Elsevier Ltd., 2016, http://dx.doi.org/10.1016/B978-0-08-100596-5.00100-1. [39] L. Barba, G. Arrighetti, S. Calligaris, Crystallization and melting properties of extra virgin olive oil studied by synchrotron XRD and DSC, Eur. J. Lipid Sci. Technol. 115 (2013) 322–329, http://dx.doi.org/10.1002/ejlt.201200259. [40] A.M.M. Guedes, R. Antoniassi, M.C. Galdeano, R. Grimaldi, M.G. De Carvalho, A.E. Wilhelm, A.G. Marangoni, Length-scale specific crystalline structural changes induced by molecular randomization of pequi oil, J. Oleo Sci. 66 (2017) 469–478, http://dx.doi.org/10.5650/jos.ess16192. [41] S. Samaram, H. Mirhosseini, C.P. Tan, H.M. Ghazali, Ultrasound-assisted extraction and solvent extraction of papaya seed oil: crystallization and thermal behavior, saturation degree, color and oxidative stability, Ind. Crops Prod. 52 (2014) 702–708, http://dx.doi.org/10.1016/j.indcrop.2013.11.047. [42] L.J. Pham, P.J. Pham, Biocatalyzed production of structured olive oil triacylglycerols, in: D. Boskou (Ed.), Olive Oil – Const. Qual. Heal. Prop. Bioconversions, InTech, 2012, pp. 447–456 http://www.intechopen.com/books/olive-oilconstituents-quality-health-properties-and-bioconversions/biocatalyzedproduction-of-structured-olive-oil-triacylglycerols. [43] T. Karupaiah, K. Sundram, Effects of stereospecific positioning of fatty acids in triacylglycerol structures in native and randomized fats: a review of their nutritional implications, Nutr. Metab. (Lond) 4 (2007) 16, http://dx.doi.org/10.1186/ 1743-7075-4-16.

temperature and pressure in subcritical extraction with n-propane, Ind. Crops Prod. 87 (2016) 64–70, http://dx.doi.org/10.1016/j.indcrop.2016.04.029. C.M. Da Silva, A.B. Zanqui, A.H.P. Souza, A.K. Gohara, S.T.M. Gomes, E.A. Da Silva, L.C. Filho, M. Matsushita, Extraction of oil and bioactive compounds from Araucaria angustifolia (Bertol.) Kuntze using subcritical n-propane and organic solvents, J. Supercrit. Fluids 112 (2016) 14–21, http://dx.doi.org/10.1016/j.supflu. 2016.02.003. L. dos S. Freitas, J.V. de Oliveira, C. Dariva, R.A. Jacques, E.B. Caramão, Extraction of grape seed oil using compressed carbon dioxide and propane: extraction yields and characterization of free glycerol compounds, J. Agric. Food Chem. 56 (2008) 2558–2564, http://dx.doi.org/10.1021/jf0732096. G. Nimet, E.A. da Silva, F. Palú, C. Dariva, L. dos S. Freitas, A.M. Neto, L.C. Filho, Extraction of sunflower (Heliantus annuus L.) oil with supercritical CO2 and subcritical propane: experimental and modeling, Chem. Eng. J. 168 (2011) 262–268, http://dx.doi.org/10.1016/j.cej.2010.12.088. I.M.M. de Carvalho, L.D. Queirós, L.F. Brito, F.A. Santos, A.V.M. Bandeira, A.L. de Souza, J.H. de Queiroz, Caracterização química da castanha de sapucaia (Lecythis pisonis Cambess.) da região da zona da mata Mineira, Biosci. J. 28 (2012) 971–977. G.L. Teixeira, S. Ávila, M. Ribani, J.L.M. Silveira, R.H. Ribani, Chemical, thermal and rheological properties and stability of sapucaia (Lecythis pisonis) nut oils: a potential source of vegetable oil in industry, J. Therm. Anal. Calorim. (2017), http://dx.doi.org/10.1007/s10973-017-6742-1 (In Proof). AOCS, Official Methods and Recommended Practices of the American Oil Chemists’ Society, 5th ed., AOCS Press, Champaign, Illinois, 1997. W.W. Christie, A simple procedure for rapid transmethylation of glycerolipids and cholesteryl esters, J. Lipid Res. 23 (1982) 1072–1075. D. Sparks, R. Hernandez, M. Zappi, D. Blackwell, T. Fleming, Extraction of rice bran oil using supercritical carbon dioxide and propane, JAOCS J. Am. Oil Chem. Soc. 83 (2006) 885–891. H. Sovová, Apparent solubility of natural products extracted with near-critical carbon dioxide, Am. J. Anal. Chem. 3 (2012) 958–965, http://dx.doi.org/10.4236/ ajac.2012.312A127. G. Knothe, Structure indices in FA chemistry. How relevant is the iodine value? J. Am. Oil Chem. Soc. 79 (2002) 847–854, http://dx.doi.org/10.1007/s11746-0020569-4. V.N. Castelo-Branco, I. Santana, V.O. Di-Sarli, S.P. Freitas, A.G. Torres, Antioxidant capacity is a surrogate measure of the quality and stability of vegetable oils, Eur. J. Lipid Sci. Technol. 118 (2016) 224–235, http://dx.doi.org/10.1002/ejlt. 201400299. M. Funasaki, R.S. Oliveira, S.P. Zanotto, C.R.F. Carioca, R.C. Simas, M.N. Eberlin, R.M. Alberici, Brazil nut oil: quality control via triacylglycerol pro fi les provided by easy ambient sonic-spray ionization mass spectrometry, J. Agric. Food Chem. 60

132