Optimized ultrasonic‐assisted extraction of papaya seed oil from Hainan/Eksotika variety

Abstract Hainan/Eksotika papaya is a popular cultivated plant in Hainan Island, China. Papaya seed oil (PSO) contains functional compounds with good antioxidant activity, especially monounsaturated fatty acids. In this work, the ultrasound‐assisted extraction (UAE) of PSO was optimized using response surface methodology. It was found that the optimal extraction performance was realized when the elevated time was set to 20 min, the ultrasound power was set to 250 W, and the n‐hexane‐to‐sample ratio was set to 16:1 (v/w). The highest yield of PSO (32.27%) was obtained under the optimal conditions, and PSO showed good oxidative stability. Differential scanning calorimetry analysis showed that the melting point of Hainan/Eksotika PSO was low, while its crystallization temperature was high. FTIR and NMR were used to analyze the chemical structure of PSO, which also proved that PSO possessed good stability without oxidative degradation. In addition, scanning electron micrograph was employed to investigate the change in seed microscopic structure. The results showed UAE caused serious structural damage of sample cell membranes and walls, which help oil access to the solvent with a high extraction ratio. The results indicated that UAE is an efficient environmental‐friendly, and promising technique could be applied to produce PSO or other edible oil with a better health‐beneficial value in food industry.

of monounsaturated fatty acids and functional phytochemicals, such as tocopherol, carotenoid, and phenolics. Moreover, papaya seed oil (PSO) was found to be robust against oxidation, which can be processed into a new type of cooking oil containing a better health-beneficial value in food industry (Samaram, Mirhosseini, Tan, & Ghazali, 2014). This provides insight into reducing environmental pollution and making waste seeds profitable. Therefore, more and more attentions have been diverted to the PSO.
Normally, the output and quality of extracted PSO are significantly associated with the extraction method and extraction condition. Soxhlet solvent extraction is a common method to extract oil from plant sources, but the process of this method is relatively time consuming and the recovery of oil is limited. Therefore, many efforts have been made to develop methods integrating thermal and mechanical treatments to enhance PSO extraction. Among them, ultrasound-assisted extraction (UAE) has been proved to be an efficient and economic "green" technology for the extraction of oil, with high oil yields, short extraction time, and a low volume of solvent (Tavares, Massa, Gonçalves, Silva, & Santos, 2017). In addition, UAE makes it possible to extract a high rate of bioactive compounds at relatively lower temperature, resulting in higher quantity and quality of final products than Soxhlet extraction (SE; Miklavcic, 2017).
The effects of temperature, time, solvent-solid ratio, ultrasound power on the extraction of oil from Batek Batu, and Sekaki were investigated. Most studies simply evaluated the yield and chemical composition of PSO treated with ultrasound, and very few studies focused on the effects of UAE on physicochemical characteristics of PSO. Especially, there is no report on the changes in seed microscopic structures before and after UAE extraction.
In this study, Soxhlet and UAE were used to extract PSO from Hainan/Eksotika papaya, and the two extraction methods were compared. To optimize the UAE of PSO from Hainan/Eksotika variety, response surface optimization was adopted using n-hexane as a solvent. This study investigated the yield and physicochemical characteristics of PSO extracted from Hainan/Eksotika variety by UAE. Furthermore, the mechanism of UAE was investigated using scanning electron microscopy (SEM). This study provides an environmentally friendly extraction method and develops a new edible oil resource, which has positive effects on the food industry.

| Plant materials and chemical reagents
Hainan/Eksotika papaya was purchased from a local market in Haikou, Hainan, China. Seeds of papaya fruit were collected, washed and dried at 60°C for 2 days, and finally ground into powder and sieved via a 40-mesh screen (Samaram et al., 2015). Standards including nonanal. octanal, 2-decenal, decanal, and 2-undecenal were purchased from Tokyo Kasei Kogyo Co., Ltd. Isopropanol and acetonitrile for HPLC were chromatographically pure. All the other chemicals and solvents were of analytical grade.

| Soxhlet extraction
Seed powder (5 g) was extracted and refluxed using a soxhlet extractor with 200 ml of n-hexane (60°C) for 10 hr at a mass-to-solvent ratio of 1:40 (w/v), resulting in seed oil according to Bhutada, Jadhav, Pinjari, Nemade, and Jain (2016) with some modification.
The resulted seed oil was collected by rotary vacuum evaporator at 60°C and then stored at 4°C.
The extraction yield can be obtained by the following equation:

| Experimental design of UAE
Before entering overall implementation of Box-Behnken central composite design, single-factor experiments were carried out in an indirect ultrasound bath. The variables evaluated in this study were as follows: solvent (acetone, ethyl acetate, n-hexane, petroleum ether), ultrasound power (150-250 W), time (5-50 min), and solventsample ratio (6:1-16:1, v/w), and these variables were investigated.
The effects of ultrasound power, extraction time, and solventsample ratio on the PSO yield were, respectively, analyzed using RSM. The single-factor experiment was carried out to determine the contribution of each variable, as shown in Table 1. The optimal conditions were obtained using Box-Behnken Design (BBD). The average yield (%) of oil was regarded as response value. Table 1 shows operating parameters suggested by SAS 8.0 (Statistical Analysis (1) Extraction yield (%) = extracted oil (g)/initial seed power (g) × 100.
TA B L E 1 Box-Behnken design of process variables along with experimental values for the yield response System Institute Inc.). According to the following quadratic polynomial regression equation, predicted yield of PSO was obtained.
where Y represents the predicted response; β 0 , β i , β ii , and β ij are regression coefficients of the intercept, linear, quadratic, and interaction terms, respectively; and X i and X j are independent variables (Wang, Wang, Wang, Xiao, & Liu, 2016).

| Physicochemical characteristics of PSO
The acid value (AV), free fatty acid contents, iodine value, p-ani-

| DSC-based thermal analysis
A Perkin-Elmer DSC with a data station was used for thermal analysis. The purge gas was nitrogen (99.99% purity). Melted sample (3 g) was added in a DSC pan and then hermetically sealed (Yanty, Marikkar, Nusantoro, Long, & Ghazali, 2014 where k is the kinetic rate constant (per hour), and R is the molar gas constant (8.314 J/mol K).

| FTIR analysis of PSO
The FTIR spectra of PSO were recorded using a FTIR spectrophotometer (Nicolet 6700; Thermo scientific) to reveal the functional groups by recording 45 scans in transmittance mode in the range of 4,000-400/cm (Bhutada et al., 2016).

| Proton nuclear magnetic resonance ( 1 HNMR) analysis of PSO
The molecular structure of the PSO extracted by UAE was analyzed using 1 HNMR. The oil sample was dissolved in the deuterated chlo-

| SEM analysis
The morphological changes in seed samples before and after UAE were determined using a Quanta-200 environmental scanning electron microscope system (FEI Company). Samples were fixed and then sputtered with a thin layer of gold. Determination was operated at an accelerating voltage of 12.5 kV under high vacuum condition (1,000 × magnification) (Jiao et al., 2014). Cao et al. (2014)

| Statistical analysis
All experiments were replicated three times. Data were expressed in the form of means ± standard deviation (SD). The data were analyzed by the IBM SPSS statistical software (SPSS Inc.). The difference was of statistical significance at p < 0.05.

| Optimization of UAE process
The results of single-factor experiments indicated that the extraction yield of PSO was associated with by solvent, ultrasound power, extraction time, and solvent-sample ratio. The highest yield was obtained when using n-hexane as extract solvent. The maximal UAE yield of PSO could be obtained by optimizing the conditions. Table S1 and Figure S1 showed that the regression model was precise. (2) There were significant differences (p < 0.05 or p < 0.01) in the linear terms (X 1 and X 3 ) and a quadratic term (X 3 2 ), while there were no significant differences (p > 0.05) in one linear term (X 2 ), three interaction terms (X 1 X 2 , X 1 X 3 , and X 2 X 3 ), and two quadratic terms (X 2 1 and X 2 2 ). The oil yield calculated can be expressed using the second order polynomial model: where Y represents the yield (%), X 1 stands for the ultrasonic power (W), X 2 represents the extraction time (min), and X 3 represents the solvent-sample ratio.
The influences of three extraction parameters can be ranked as follows: ultrasonic power > solvent-sample ratio > ultrasonic time.
The ultrasonic power showed the greatest influence on the PSO yield in the UAE, and this effect was clearly observed when comparing the results of experiments 1 and 3 in Table 1. Tavares et al. (2017) found that an increase in the variables time, temperature, and solvent: seed ratio favored the removal of oil from the crambe seeds. In addition, significant higher PSO yields were obtained when the solvent-to-solid ratio increased, due to a higher driving force and a less viscosity in the more diluted solution (de Mello, dos Santos Garcia,

& da Silva, 2017). A literature reported by Stevanato and da Silva
(2019) evaluated the extraction of radish seed oil at 45°C and 60 min and observed that increasing the solvent-to-seed ratio from 4:1 to 12:1 increased the oil yield from 17.10% to 23.22%, respectively.
Another study reported an increase in the yield of macauba pulp oil (28.15%-30.70%) when increasing the relative amount of solvent  (Samaram et al., 2015).
The optimal extraction performance could be obtained when ultrasonic power was 249.59 W, extraction time was 20.34 min, and solvent-sample ratio was set to 16.22:1 (v/w) (Fig. S1), which is in well consistence with the results of others using the response surface (Jadhav, Holkar, Goswami, Pandit, & Pinjari, 2016;Li, Lu, et al., 2012;Li, Qu, Zhang, & Wang, 2012). In view of the actual operation, ultrasonic power was set to 250 W, extraction time was set to 20 min, and solvent-sample ratio was set to 16:1, respectively. The actual PSO yield obtained under optimum conditions was 32.27 ± 0.39%, which was very close to the predicted value of 32.55% of the regression model. This suggests that the response model could predict and evaluate the oil yields from papaya seed.  It was found that the specific gravity, refractive index, PV, AV, iodine value, TV, saponification value, and the content of free fatty acid of the oil extracted by UAE were significantly lower than those extracted by SE (p < 0.05). However, the extraction methods did not make significant impact (p > 0.05) on the fatty acid composition of PSO, or on the fatty acid compositions of sour cherry kernel oil (Samaram et al., 2014) and PSO (Phan, Junyusen, Liplap, & Junyusen, 2018).   extraction nor aqueous enzymatic extraction made significant effects on thermal behavior of PSO (Delfanian, Esmaeilzadeh Kenari, & Sahari, 2015).

| Thermal behavior
From the DSC curves of oxidations of blackberry and raspberry seed oil, three characteristic points were observed, including onset temperature (T on ), the first peak (T p1 ), and the second peak (T p2 ).
Among them, onset temperature (T on ) and the first peak (T p1 ) were clearly observed in the DSC curve of the PSO oxidation. Relevant study showed when assessing the oil oxidative stability by the nonisothermal DSC method, only the onset of oxidation (T on ) and the first peak (T p1 ) should be considered (Zhang et al., 2018). Therefore, this method was suitable for calculating activation energy and the pre-exponential factor for the peak temperature (T p1 ). Table 3 shows the activation energy (E a ) and the pre-exponential factor obtained by the Kissinger method. T p1 fitted well to the experimental results for both samples (R 2 > 0.90). The E a of UAE-PSO (100.32 kJ/ mol) was higher than that of SE-PSO (90.51 kJ/mol), which correlated well with values reported previously for related oils, such as sunflower (90.74 kJ/mol) and sesame oils (93.55 kJ/mol), canola oil (89.94 kJ/mol), soybean oil (92.42 kJ/mol), corn oil (88.14 kJ/mol), and olive oil (86.86 kJ/mol) (Farhoosh, Niazmand, Rezaei, & Sarabi, 2008;Ghosh, Upadhyay, Mahato, & Mishra, 2019). Higher E a values indicate a slower rate of lipid oxidation (Tan, Man, Selamat, & Yusoff, 2002); hence, UAE-PSO proposed higher oxidative stability than SE-PSO and most vegetable oils. Figure 2a shows the broad peak at 3,467/cm (O-H stretch) and

| FTIR analysis of PSO
723/cm of ortho-substituents. The peaks at 2,924/cm and 2,853/ cm were, respectively, ascribed to symmetrical and asymmetrical stretching of the C-H in CH 2 or CH 3 group in fatty acids. The peak at 1,760-1,665/cm was attributed to the ester functional group. The peak at 3,000-2,700/cm was due to the carboxyl group in the free fatty acid. The peaks at 3,010-3,050/cm and 1,600-1,450/cm were due to the C=C bonds on benzene (Bhutada et al., 2016). Additionally, remarkable peak 2,086/cm (N=C=S) was observed, which indicated that the presence of benzyl was othiocyanate in UAE-PSO. In contrast, this compound was not found in SE-PSO.

| 1 HNMR analysis of PSO
The 1 HNMR spectra of POS extracted by SE and UAE were shown in (f) and 4.28 ppm (g) were ascribed to glyceryl methylenes. The multiplets at 5.25 ppm (h) and 5.30 ppm were ascribed to glyceryl methines and olefinic protons, respectively (Timilsena et al., 2017).
After oxidation of edible oils, the oxidation products showed signals at 8.09-8.19 ppm for hydroperoxides proton and 9.30-9.90 ppm for aldehydes (Bhutada et al., 2016). Here, no hydroperoxide and aldehyde signals were found in this study, indicating that no oxidative degradation took place during SE and UAE.

| Analysis of aldehyde composition of PSO
As the main secondary oxidation products, aldehydes were taken as oxidation indicators. The content of oleic acid in PSO from Malaysia/ F I G U R E 2 FTIR (a) and 1 H NMR (b) spectra of papaya seed oil (PSO) from Soxhlet extraction (SE) and ultrasoundassisted extraction (UAE) PSO were lower than those in SE-PSO ( Figure 4). This suggested that the oxidation stability of UAE-PSO was higher than SE-PSO.

ACK N OWLED G M ENTS
This study was supported by National Natural Science Foundation of China (Grant Nos. 31660495;31801494).

CO N FLI C T O F I NTE R E S T
There is no conflict of interests regarding the publication of this paper.

E TH I C A L S TATEM ENT
This study does not involve any human or animal testing.