Optimisation of Microwave-Assisted Extraction of Pomegranate ( Punica granatum L.) Seed Oil and Evaluation of Its Physicochemical and Bioactive Properties

,


Introduction
Microwave-assi sted solvent extraction (MASE) is a process that has emerged as an att ractive alternative oil extraction method in recent years.The rapid heating and destruction of biological cell structure in a microwave provide more eff ective extraction in shorter time than conventional processes.Moreover, MASE requires small amount of solvent for extraction and produces high-quality oil regarding chemical and physical properties.Another important advantage of this process is the lower energy requirement resulting in a signifi cant decrease in environmental impact and fi nancial costs (1)(2)(3)(4)(5)(6)(7)(8).
In the last decade, researchers have started to focus on the microwave-assisted extraction instead of the conventional methods for the extraction of natural com-pounds such as pectin, essential oil and phenolics.MASE has been extensively studied to investigate its impact on the extraction of high-value components with high yield and quality from various plant food materials, industrial food wastes and by-products.It has been widely used for the extraction of lycopene from tomato peel (9), polyphenols from red grape pomace, grape seed and potato peel (10)(11)(12), or essential oil and pectin from orange peel (13).(15,16).It contains tocopherols, phytosterols and punicic acid, which have several potential health benefi ts (17).Punicic acid constitutes 74-85 % of the total fatt y acid content of pomegranate seed oil and is known for its antioxidant, antitumour, immunomodulatory, anti-atherosclerotic and serum lipid-lowering activities (18,19).Eikani et al. (20) compared the effi ciency of Soxhlet, cold pressing and superheated hexane extraction methods in the extraction of pomegranate seed oil.Tian et al. (21) optimised the conditions for ultrasonic-assisted extraction of pomegranate seed oil.Fadavi et al. (22) investigated the total lipid content and fatt y acid composition of pomegranate seed oil extracted from 25 diff erent varieties grown in Iran.Sadeghi (23) evaluated the physical and chemical characteristics of four pomegranate cultivars.However, extraction of pomegranate seed oil has not been evaluated previously in a closed-vessel high-pressure microwave extraction system.The objectives of the presented study are to observe the eff ects of extraction time, solvent-to-solid ratio (by mass), particle size and microwave power on oil extraction yield in a microwave system using response surface methodology and to compare the yield and the quality parameters such as physical, chemical and bioactive properties of the oil obtained by MASE and cold solvent extraction.

Sample preparation
The seeds were dried to moisture content of 3.5 % in a vacuum oven (model VD 23; Binder Inc., Bohemia, NY, USA) at 35 °C for 3 h.Dried seeds were ground in a coff ee grinder (model PRG 259; Premier, Istanbul, Turkey).Ground particles were sieved through meshes into diff erent particle sizes: fi ne particles d 1 =0.125-0.450mm (1), moderately coarse particles d 2 =0.450-0.530mm (2) and coarse particles d 3 =0.530-0.800mm (3), collected and used in the experiments.Ground particles were sealed in glass bottles and kept at 4 °C until extractions were performed.

Conventional extractions
Cold solvent and Soxhlet extraction methods were performed using the method proposed by Abbasi et al. (2) and Jiao et al. (24) with some modifi cations to compare their oil extraction effi ciencies with that of microwave-assisted solvent extraction (MASE).Conventional Soxhlet extraction was performed using 10 g of fi ne powdered seeds (d 1 =0.125-0.450mm) and 220 mL of n-hexane in a classical Soxhlet apparatus at 110 °C for 8 h.For cold solvent extraction, 40 g of ground seeds and 400 mL of n--hexane were put into a glass beaker and stirred by magnetic stirrer (model 613.01.001;Isolab, Werthelm, Ger m-any) with extraction time of 8 h at 25 °C.Finally, the solid residue was precipitated by centrifugation (3461×g, 10 min, EBA 20; Hett ich, Tutt lingen, Germany).The supernatant was separated by decantation.Hexane was removed at 40 °C using rotary vacuum evaporator (Hei-VAP Advantage HL/G1; Heidolph Instrument GmbH & Co. KG, Schwabach, Germany) in both extraction methods.The extracted oil was stored at -20 o C until the analyses were carried out.The yield was determined using the following equation:

Microwave-assisted solvent extraction
Microwave-assisted solvent extraction (MASE) was performed in a CEM Discover SP-D microwave reactor at 2450 MHz (CEM Corporation, Matt hews, NC, USA).The extraction was checked and monitored via computer.Ground pomegranate seeds and n-hexane at diff erent solvent-to-sample ratios (Table 1) were put into 35-mL microwave quartz vessel closed with snap-on caps.The values of power and time were adjusted as given in the experimental central composite design (Table 1) using CEM Synergy soft ware (CEM Corporation).Stirring was set at high level.The extractions were performed in dynamic mode (maximum power P=300 W) and PowerMax function, which blows air to eliminate additional heating, was on.The required time for heating up and cooling down were not included in total extraction time.Aft er extraction, the solid residue was precipitated by centrifugation (3461×g, 10 min) and supernatant was collected.Hexane was removed at 40 °C using rotary vacuum evaporator (Hei-VAP Advantage HL/G1; Heidolph Instrument GmbH & Co. KG).The extracted oil was stored at -20 °C prior to analysis.The yield was calculated using Eq. 1.

Experimental design and optimisation by response surface methodology
A three-level, four-factorial face-centred central composite design was applied to evaluate the eff ects of extraction parameters on oil extraction yield and to determine optimum extraction conditions for obtaining highest extraction yield.The design consisted of 30 experimental runs with six replications at the central point (Table 1).The extraction variables were time (5-20 min), power (176-300 W), solvent-to-sample ratio (2:1, 6:1 and 10:1) and particle size (d=0.125-0.800mm).The response was the oil extraction yield (Y).The data were analysed as reported by Keskin et al. (25).The predicted values given by the model fi tt ing technique in Design Expert v. 7.0 (Stat--Ease, Inc., Minneapolis, MN, USA) were closely correlated with the experimental values.

Determination of fatt y acid composition
The fatt y acid composition of the extracted oil was determined by using Agilent 7890A gas chromatograph (Agilent Technologies, Santa Clara, CA, USA) with a split/ splitless injector, fl ame ionisation detector and HP-88 capillary column (88 % cianopropylaryl; 100 m×0.250 mm, 0.20 μm i.d.).The method proposed by Çift çi et al. (26 was used with some modifi cations.The injector and detector temperatures were 250 and 260 °C, respectively.The oven temperature was scheduled as follows: 1 min at 120 °C, from 120 to 175 °C at 10 °C/min, 10 min at 175 °C, from 175 to 210 °C at 5 °C/min, 5 min at 210 °C, from 210 to 230 °C at 5 °C/min and 5 min at 230 °C.Helium was the carrier gas with a fl ow rate of 1.5 mL/min.

Determination of free fatt y acid and total phenolic content, and peroxide value
The free fatt y acid content and peroxide value of oil were determined according to AOCS Ca 5a-40 (27) and Cd 8-53 ( 28) methods, respectively.The total phenolic content of the extracted oil was determined using the Folin-Ciocalteu colourimetric method proposed by Gutfi nger (29) with some modifi cations.The oil obtained by microwave-assisted or cold solvent extraction (2.5 g) was dissolved in n-hexane (5 mL).Oil-in-hexane solution together with 5 mL of methanol and water (60:40, by volume) mixture was vortexed to extract the phenolic compounds.Centrifugation at 3461×g for 10 min (EBA 20; Hett ich) was used to separate the two phases.This extraction procedure was done in triplicate.The extracts in methanol were mixed and 0.2 mL of the methanolic phase was diluted to 5 mL with distilled water.Folin-Ciocalteu reagent (0.5 mL) was added to this mixture.Aft er 3 min, 1 mL of Na 2 CO 3 (20 %, by mass per volume) was added to the reaction mixture, which was diluted to a fi nal volume of 10 mL with distilled water and held for 1 h in the dark.The absorbance values were measured against a blank sample at 765 nm using Lambda 25 UV/Vis spectrophotometer (PerkinElmer, Shelton, CT, USA).The calibration curve was obtained using gallic acid standard solutions (0-60 mg /mL).The results were expressed in mg of gallic acid equivalents per g of sample dry mass.

Antioxidant activity assay
The antioxidant activity of the oil was determined using 2,2-diphenyl-1-picrylhydrazyl (DPPH) free radical with the method proposed by Kalantzakis et al. (30).Briefly, 4 mL of freshly prepared DPPH solution (0.1 mM) were added to 1 mL of oil and ethyl acetate solution at diff erent concentrations (0.05-25 mg/mL).Aft er 30 min of incubation at 25 °C in the dark, the absorbance of the fi nal solution was measured at 515 nm with Lambda 25 UV/Vis spectrophotometer (PerkinElmer).The percentage of inhibition was calculated using the following equation: where A c and A s are the absorbance of the control and sample at 515 nm, respectively.DPPH scavenging activity was expressed as IC 50 , which indicated the eff ective sample concentration needed to scavenge 50 % of DPPH radicals and was calculated by a linear regression analysis between the oil concentration and the percentage of inhibition.

Scanning electron microscopy analysis
Untreated pomagranate seed and solid residues aft er conventional and microwave-assisted extraction were examined using scanning electron microscopy (SEM) to analyse the eff ect of extraction methods on the surface morphology of the seeds.Gold/palladium was used for the coating of the samples in an SC7620 sputt er coater (Emitech, Kent, UK).Images of samples were taken with JSM-6390LV (JEOL Ltd., Tokyo, Japan) scanning electron microscope.The SEM images were obtained at 12.5 kV under high vacuum condition and 1000× magnifi cation.

Statistical analysis
The independent t-test was used to evaluate diff erences in the properties of the oil samples obtained by microwave-assisted and cold solvent extraction.The data were analysed using the SPSS v. 22.0 statistical soft ware (SPSS Inc., Chicago, IL, USA).

Comparison of extraction effi ciencies
Microwave-assisted solvent extraction gave a higher extraction yield of 35.10 % in 5 min than those of Soxhlet extraction (34.70 % in 8 h) and cold solvent extraction (17.50 % in 8 h).Orak et al. (15) and Özgul-Yücel (16) found that the oil content of seeds from diff erent genotypes of Punica granatum L. was between 17.84 and 24.96 % (on dry mass basis), which is lower than our results.The diff erences in the yields might be a result of the genetic backgrounds and the growth conditions of the pomegranate or the applied extraction methods.Taghvaei et al. (6) reported that the eff ectiveness of microwaves on rapid extraction of oil from oilseeds was related to the destruction of oil cell structures within the plant tissues through denaturation of cell proteins by heat generated from the movement of polar molecules including water.

Eff ects of microwave extraction parameters
The eff ects of diff erent parameters on the MASE of oil from pomegranate seeds were analysed.The experimental and predicted values of the oil extraction yield at each of the 30 runs given by response surface methodology (RSM) are shown in Table 1.

Particle size
Particle size was the most important factor that affected the oil yield in MASE (Table 2).Negative coeffi cient of the term stated that decreasing the particle size increased the oil extraction yield.The smaller particle size might increase the penetration of irradiation into the interior cell walls and the mass transfer surface area by association of the seed matrix and hexane enhancing the oil extraction.Kwon et al. (31) and Singh et al. (32) also documented that there was an inverse relationship between the particle size and microwave extraction yield of diff erent compounds.The eff ect of interaction between particle size and time on oil extraction yield was also statistically important (Table 2).Fig. 1a illustrates the interaction effect of the particle size and time on extraction yield at a constant solvent-to-sample ratio of 6:1 (by mass) and power of 238 W. The yield at the end of 5-minute extraction was signifi cantly higher (32 %) when using fi ne particles than when coarse particles were used (11 %) under the same extraction conditions.The yields were 34 and 11 % at the end of 20 min of extraction when fi ne and coarse particles were used, respectively.This inverse eff ect might be explained with limited penetration of microwave irradiation into the coarse seeds.It is basically because of zero dielectric constant of cellulose, which is the main component of the seed.Thus, extraction yield could not be improved when using larger particles even though the time increased.Quadratic term of particle size was negative, indicating the presence of maximum value for this variable (Table 2).

Solvent-to-sample ratio
Statistical results showed that the solvent-to-sample ratio aff ected extraction yield signifi cantly (Table 2).The solvent-to-sample ratio had statistically signifi cant interaction eff ect with particle size.When fi ne particles of pomegranate seed were used in the microwave extraction, increasing solvent-to-sample ratio from 2:1 to 10:1 (by mass) caused an increase in the extraction yield from 29 to 36 %.However, when the extraction was done with coarse particles of pomegranate seed, increasing solvent--to-sample ratio from 2:1 to 10:1 (by mass) increased the extraction yield from 8 to 12 % (Fig. 1b).This might be explained again by the limited penetration of microwaves into the interior of the coarse seeds.Because of the limited penetration, all of the oil could not be released from the seeds.Hence, there was no extra oil to be dissolved in the medium even though the amount of solvent increased.Quadratic term of solvent-to-sample ratio was found to be statistically signifi cant (Table 2).

Time
The linear term of extraction time was statistically ineff ective on oil extraction yield (Table 2).Nde et al. (33) also reported that time was statistically ineffi cient on the microwave extraction of neem oil.However, its eff ect can be understood more obviously by considering the principle of closed-vessel high-pressure microwave extraction.In the present study, closed system microwave application resulted in high pressure extraction of the oil.The combination of the microwaves and high pressure resulted in higher extraction yields at short extraction times  (34) were 66 and 83 min, respectively, which was longer than the value found in this study.In Figs.1c and  d, it can also be seen that increasing the extraction time increased the oil extraction yield.However, these improvements were not statistically signifi cant (Table 2).
Similar to this result, Wang et al. (35) also found that the increase in the extraction yield was very small as a function of time in microwave extraction.The reason for the insignifi cant increase could be explained by the fact that MASE of diff erent compounds is usually completed within a few minutes to half an hour, depending on the extracted material as mentioned in literature (33,(36)(37)(38).

Power
The linear term of power was omitt ed by backward elimination to conserve the hierarchy of the model since it had statistically insignifi cant eff ect on the oil extraction yield.Figs.1d, e and f show that increasing the power from 176 to 300 W did not improve the oil extraction yield.The reason for this result might be related to the limited increase in extraction temperature with increasing power.It is well known that higher temperatures increase the oil extraction effi ciency due to the higher solubility of oil at higher temperatures.Hexane, which is a microwave-transparent solvent and has low dielectric constant, cannot be heated up to temperature higher than 55 °C during extraction.Thus, hexane might limit the capacity of power to increase extraction effi ciency.

Optimisation and validation of extraction conditions
Extraction conditions were optimised for the highest oil extraction yield using Design Expert v. 7.0 soft ware (Stat-Ease, Inc.).The predicted oil extraction yield was 35.19 % under the optimum conditions of 220 W, 5 min, solvent/sample ratio 10:1 (by mass) and d 1 =0.125-0.450mm (fi ne particles, size 1).Three experiments were performed at these optimum conditions to validate the predicted result, and average oil extraction yield was 34.91 %.The predicted and actual values were in good agreement.

Fatt y acid composition
The pomegranate seed oil samples obtained from microwave-assisted and cold solvent extraction had almost identical fatt y acid profi les (Table 3).The total unsaturat-ed fatt y acid content of pomegranate seed oil obtained by MASE was 95.57%.The predominant fatt y acid was punicic acid (86.53 %) in the oil extracted using microwaves (Table 3).Oleic, linoleic, palmitic and stearic acids were present in minor amounts at 4.10, 3.84, 2.04 and 1.71 %, respectively.The punicic acid and total unsaturated fatt y acid content of the investigated pomegranate seed oil were higher than in the papers of Pereira de Melo et al. (14), Fadavi et al. (22) and Khoddami et al. (39).The dissimilarity might be related to the diff erences in pomegranate cultivars and climatic conditions during growth.

Peroxide value and acidity
Peroxide value and acidity are the indices for the determination of the oxidative degradation of products in oil.The peroxide value of pomegranate seed oil obtained by cold solvent extraction was 4 mmol of O 2 per kg, while that of the oil extracted by microwave-assisted extraction was 0 mmol of O 2 per kg (Table 4).The long extraction time (8 h) and an extraction process under atmospheric pressure could be the reasons for higher peroxide value of the oil extracted by cold solvent extraction.In the oil extracted by cold solvent and microwave-assisted solvent extractions, there was 0.42 and 0.44 % free fatt y acids, expressed as punicic acid, respectively.

Total phenolic content
The total phenolic content expressed in gallic acid equivalents of the pomegranate seed oil obtained by microwave-assisted and cold solvent extraction were 7.42 and 1.73 mg/g, respectively (Table 4).The higher total phenolic content of pomegranate seed oil obtained by MASE can be related to higher pressure during the extraction.The presented extraction system combines the advantage of microwave-assisted and pressurised solvent extraction.Bayramoglu et al. (40) also reported that higher internal pressure of the solid media and hence enhancement of the extraction may be the reason for higher Antioxidant activity IC 50 value of the oil extracted by microwave-assisted extraction (5.12 mg/mL) was signifi cantly lower than that of the oil extracted by cold extraction (17.00 mg/mL) (Table 4).This signifi cant diff erence in the IC 50 values might be related to the diff erences in total phenolic content of the extracted oil depending on the extraction method.He et al. (42) demonstrated a signifi cant relationship between DPPH activities and total phenolics (R 2 =0.751) in pomegranate seed residues.Gil et al. (43) also reported that pomegranate fruit is a rich source of two types of polyphenolic compounds: anthocyanins and hydrolysable tan nins, which account for 92 % of the antioxidant activity of the whole fruit.This showed that higher total phenolic content in the oil obtained by microwave-assisted extraction signifi cantly increased the antioxidant activity of the oil when compared to that of the oil obtained by cold solvent extraction.

Colour
The colour values of the oil extracted by two diff erent methods were signifi cantly diff erent (Table 4).The lightness (L*) of pomegranate seed oil extracted by cold extraction method (58.91) was higher than that of the oil extracted by MASE (56.03).The a* value measures redness (+) and greenness (-) and the b* value indicates yellowness (+) and blueness (-).The b* value of the oil extracted by MASE (22.06) was higher than that of the oil extracted by cold extraction (14.44), while a* value of the oil extract-ed by MASE (-5.64) was lower than that of the oil extracted by cold solvent extraction (-2.45).These results showed that MASE was more effi cient in extracting the chlorophyll and carotene present in the pomegranate seeds.
Structural changes of pomegranate seeds SEM analyses were performed to observe the microscopic changes in pomegranate seed before and aft er extraction to compare the infl uence of conventional and microwave-assisted solvent extractions on the pomegranate seed structure.The fat globules were dispersed uniformly in the tissues of pomegranate seed before the extraction (Fig. 2a).On the SEM images of the residues aft er Soxhlet and cold solvent extraction, the fat globules disappeared while the cell structure was preserved and the cells were still entirely unbroken (Figs.2b and c).However, MASE caused some structural changes in the seed tissues (Fig. 2d).Most of the cell walls and membranes of the pomegranate seeds were ruptured and broken down completely aft er MASE (Fig. 2d).It is clearly seen that the oil can be released from the cell structure and extracted effi ciently by cell wall rupture in a short time.There was a good correspondence between these results and the fi ndings of Jiao et al. (24), who studied the Soxhlet and MASE of pumpkin seeds.

Conclusion
Closed-vessel high-pressure microwave system was optimised to obtain the highest performance in extraction of pomegranate seed oil.Verifi cation experiments resulted in the extraction yield of 34.91 % under the optimum extraction conditions and are in good agreement with the The oil extraction yield obtained by MASE was higher than those obtained by conventional extractions.The fatt y acid compositions of the oil extracted by cold and microwave-assisted solvent extractions were statistically similar to each other (p<0.05).Oil extracted by MASE had signifi cantly lower peroxide value (0 mmol of O 2 per kg of oil), free fatt y acidity (0.42 %) and higher total phenolic content expressed in gallic acid equivalents (7.42 mg/g) and antioxidant activity (5 mg/mL) than those of the oil obtained by cold extraction.SEM results demonstrated that microwave method broke down cell walls and membranes effi ciently, thus increasing the effi cient release of oil from the seed in shorter time than in conventional extraction.The results indicated that higher quality oil could be obtained using MASE than when using cold extraction.The application of high-efficiency short-time MASE can be a valuable alternative to conventional oil extraction methods, especially for healthcare, pharmaceutical and healthy food industries.

Fig. 2 .
Fig. 2. SEM images of pomegranate seed samples: a) untreated pomegranate seed, b) solid residue aft er Soxhlet extraction, c) solid residue aft er cold solvent extraction and d) solid residue aft er microwave-assisted solvent extraction (MASE).MASE was performed under optimum conditions

Table 1 .
A three-level, four factorial face-centred central composite design generated for microwave-assisted extraction of pomegranate seed oil

Table 2 .
ANOVA and model equation for response surface quadratic model of microwave-assisted oil extraction b signifi cant at 'Prob>F'<0.05 than when using domestic or focused microwave ovens.Optimum extraction times found by Jiao et al. (24) and Gai et al.

Table 3 .
Fatt y acid composition of pomegranate seed oil extracted by microwave-assisted and cold solvent extraction (41)olic content of the oil obtained by MASE.Similarly, Haddadi-Guemghar et al.(41)also found that the phenolic content of the oil extracted by conventional method was signifi cantly lower than that of the oil obtained by MASE.