Application of an aqueous enzymatic–ultrasound cavitation method for the separation of Sapium sebiferum seed kernel oil

Highlights

• •S. sebiferum seed oil was gained by aqueous enzymatic-ultrasound cavitation method.
• •The extraction conditions were optimized with mathematical design models.
• •The obtained S. sebiferum seed kernel oil is abundant in unsaturated fatty acids.
• •The proposed method yielded the oil with lower acid value to that of Soxhlet method.

Abstract

An aqueous enzymatic–ultrasound cavitation extraction (AEUCE) method was developed to separate Sapium sebiferum seed kernel oil. In this process, neutral proteinase was screened as the propriate enzyme. The Plackett–Burman and Box–Behnken designs were employed to optimize AEUCE. We determined the optimal extraction conditions, producing an oil yield of 84.22 ± 3.17 %. Gas chromatography–mass spectrometry (GC–MS) analysis indicated that the S. sebiferum seed kernel oil was abundant in unsaturated fatty acids (>92 %) and that the compositions of the fatty acid profiles extracted by AEUCE were similar to those obtained from Soxhlet extraction, but their contents were slightly different. The physicochemical properties analysis showed that the oil extracted by AEUCE was comparable to that obtained from Soxhlet extraction. The results showed that the developed AEUCE is an efficient technique that can separate high-quality plant oils. The S. sebiferum seed kernel oil obtained from this extraction method is a promising substitute for vegetable oils used in biodiesel production.

1. Introduction

The growth of the human economy and population has increased energy demands, and traditional fossil fuels have been the main source of this energy throughout history [1]. However, heavy environmental pollution and the non-renewable nature of fossil fuels have resulted in the need to identify low-carbon energy sources [2]. Biodiesel is prepared via the esterification (transesterification) of vegetable oils or animal fats, which are feedstock oils. This is an ideal alternative to fossil fuels owing to its clean, degradable, renewable, and sustainable properties [3], [4]. However, using vegetable oils to produce biodiesel threatens food security because they are not economical [5]. Therefore, using non-edible oils to produce biodiesel is crucial to balance food and energy demands.

Sapium sebiferum (in the Euphorbiaceae family), also known as the Chinese tallow tree, is distributed worldwide [6]. As one of the four woody oil crops, S. sebiferum has been cultivated as an economic crop for approximately 1,400 years in China [7], [8]. Its seeds possess more oil than cottonseed, flaxseed, and soybeans, as each seed comprises approximately 50 % oil, with the proportion of unsaturated fatty acids reaching up to 90 % [9], [10]. S. sebiferum oil has been proposed as a potential substitute for biodiesel production. Indeed, its outstanding flash point, low cold flow value, and satisfactory oxidation stability satisfy the EN 14214 and ASTM 6751–03 standards [8].

Soxhlet extraction (SE) is typically used to extract plant seed oils. However, it has many shortcomings, such as being time-consuming, incurring high costs, creating residual solvents, and causing environmental problems [11]. The mechanical press method can be employed as a potential method to avoid organic solvent consumption. However, it is also associated with low yields and accessory substances. The aqueous enzymatic method is an alternative method for separating plant oils from oil-bearing seeds, with its advantages of greenness, eco-friendliness, and the lack of need for toxic solvents [12]. Owing to the low solubility of triglycerides in aqueous media and enzyme action, the aqueous enzymatic method can efficiently extract oils from plant materials to ensure a high-quality product [13], [14]. The separation sciences have also given increasing attention to ultrasound treatment as the waves can trigger cavitation phenomena, in which the cavitation bubbles gather and burst on the solid surface at a higher amplitude of ultrasound irradiation, rupturing plant cell walls and accelerating the diffusion of target compounds into the solvent [15], [16]. In this study, we developed an aqueous enzymatic–ultrasound cavitation extraction method (AEUCE) to extract S. sebiferum seed kernel oil.

The Plackett–Burman design (PBD) and Box–Behnken design (BBD) were applied to optimize the AEUCE process. SE was also performed, and the results were compared with the developed AEUCE. The oils obtained from these two methods were analyzed using gas chromatography–mass spectrometry (GC–MS) analysis The physicochemical properties of the S. sebiferum seed kernel oil were determined, including the specific gravity, acidity, iodine, and saponification values. The morphologies of the residues extracted by these two methods were also examined using scanning electron microscopy (SEM).

2. Materials and methods

2.1. Materials and regents

S. sebiferum seeds were collected from the lawns of Jiangxi Normal University (Nanchang, China) and identified by Professor Ronggen Deng (Jiangxi Normal University, China). These seeds were divided into two parts: the wax layers and the kernels. Before extraction, the seed kernels were crushed and dried at room temperature.

Neutral proteinase, pancrelipase, α-glucosidase, and α-amylase were purchased from Shanghai Yuanye Biotechnology Co., Ltd. (Shanghai, China). In addition, pectinase, hemicellulase, cellulase, and other chemicals were purchased from Aladdin (Shanghai, China). Finally, deionized water was prepared using a Milli-Q water purification system (Millipore, Billerica, MA, USA).

2.2. Extraction of S. Sebiferum seed kernel oil

For SE, 5 g of the pulverized seed kernel powder was suspended in 75 mL of n-hexane and heated for 4 h using an electric jacket. Solvent exchanges occurred 19 times during this process. After SE, the collected suspension was evaporated using a rotating evaporator, resulting in the S. sebiferum seed kernel oil, which was then dried with anhydrous sodium sulfate. The seed kernel oil yield obtained from SE was calculated by dividing the mass of the extracted oil by the mass of the dried seed kernels, resulting in 40.60 %.

For AEUCE, 5 g of seed kernels were mixed with different enzyme solutions and then incubated in a digital water bath oscillator (Kunshan, Jiangsu). After incubation, the mixtures were placed in a KH–250DE ultrasound bath (Kunshan, Jiangsu) for ultrasound cavitation treatment, and the suspension was centrifuged for 20 min at 5000 rpm. The S. sebiferum seed kernel oil was contained in the top layer (free oil) and middle layer (a water–oil emulsification). The oil in this middle layer was separated through a circular freeze–thaw procedure, followed by centrifugation to obtain the total seed kernel oil yield. The collected oils were then combined and dried with anhydrous sodium sulfate. The seed oil yield from AEUCE is the percentage relative to that of SE. Pre-test experiments were performed to ascertain the appropriate range of the six variables in AEUCE, which were then used in the mathematical optimization conducted using the PBD and BBD. These variables were the enzyme loading amount (A), incubation time (B), water-to-solid ratio (C), ultrasound time (D), ultrasound power (E), and ultrasound temperature (F).

2.3. Experimental design

2.3.1. PBD

The PBD can simultaneously assess the effects of multiple factors at different levels, thus allowing the selection of significant factors for further optimization [17]. To estimate the effects of the extraction variables on the seed kernel oil yield, the six independent variables were analyzed to determine the significant parameters in AEUCE. As shown in Table 1, 12 runs were evaluated using a multivariate regression analysis at the low level (–1) and high level (+1). Additionally, a first-order polynomial mathematical model was applied to evaluate the influences of the six variables on oil yield, as presented in Eq. (1):

$$Y={\beta }_0+\sum {\beta }_i{X}_i+X_i$$ (1)

where Y is oil yield (%), β₀ represents the intercept term, β_i denotes the linear coefficient, and X indicates each independent parameter.

Table 1.

Plackett–Burman design results for six variables in coded and actual values.

Runa	A	B	C	D	E	F	Yield (%)
							Actual	Predicted
1	3 (+1)	2 (-1)	10 (+1)	40 (+1)	350 (-1)	60 (+1)	80.23	79.84
2	3 (+1)	4 (+1)	6 (-1)	20 (-1)	350 (-1)	60 (+1)	77.48	77.46
3	1 (-1)	4 (+1)	10 (+1)	40 (+1)	350 (-1)	40 (-1)	74.91	75.45
4	1 (-1)	2 (-1)	10 (+1)	20 (-1)	490 (+1)	60 (+1)	75.59	76.18
5	1 (-1)	4 (+1)	10 (+1)	20 (-1)	490 (+1)	60 (+1)	76.98	76.39
6	3 (+1)	2 (-1)	6 (-1)	20 (-1)	490 (+1)	40 (-1)	78.65	78.98
7	3 (+1)	4 (+1)	10 (+1)	20 (-1)	350 (-1)	40 (-1)	77.54	77.66
8	1 (-1)	2 (-1)	6 (-1)	20 (-1)	350 (-1)	40 (-1)	73.07	72.65
9	3 (+1)	4 (+1)	6 (-1)	40 (+1)	490 (+1)	60 (+1)	81.35	81.59
10	1 (-1)	4 (+1)	6 (-1)	40 (+1)	490 (+1)	40 (-1)	77.28	76.99
11	3 (+1)	2 (-1)	10 (+1)	40 (+1)	490 (+1)	40 (-1)	81.84	81.57
12	1 (-1)	2 (-1)	6 (-1)	40 (+1)	350 (-1)	60 (+1)	74.88	75.05
ANOVA
Source		Sum of squares		Degree of freedom	Mean square	F Value	P value	Inference
Model		77.92		6	12.99	38.62	0.0005	***
Residual		1.68		5	0.34
Cor Totalb		79.60		11
Regression data
Term		Effect	Coefficient	Standard error	F value	T value	P value	Inference
A		4.06	2.03	0.17	147.32	11.94	< 0.0001	***
B		0.21	0.11	0.17	0.41	0.65	0.5520	ns
C		0.73	0.36	0.17	4.75	2.12	0.0811	ns
D		1.86	0.93	0.17	30.98	5.47	0.0026	**
E		2.26	1.13	0.17	45.71	6.65	0.0011	**
F		0.54	0.27	0.17	2.57	1.59	0.1698	ns

^aA: enzyme loading amount, %; B: incubation time, h; C: water to solid ratio, mL/g; D: ultrasound time, min; E: ultrasound power, W; F: ultrasound temperature, °C.

^bCor Total: Totals of all information corrected for the mean.

*Significant at P ≤ 0.05, ** Significant at P ≤ 0.01, *** Significant at P ≤ 0.001, ^ns Not significant.

2.3.2. BBD

The BBD is a multivariate second-order factorial design that makes the analysis process more effective, simpler, and easier to comprehend in terms of the response surface methodology (RSM) [18]. The three most significant independent variables were screened in the PBD to facilitate further optimization through the BBD at three levels (low, central, and high). The AEUCE process was analyzed using multiple regression to fit the second-order polynomial model as follows:

$$Y={\beta }_0+\sum _{i=1}^3{\beta }_i{X}_i+\sum _{i=1}^3{\beta }_{\mathit{ii}}{X}_i^2+\sum _{i=1}^3\sum _{j=i+1}^3{\beta }_{\mathit{jj}}{X}_i{X}_j$$ (2)

where Y denotes the seed kernel oil yield; β₀, β_i, β_ii, and β_ij denote the constant coefficients of intercept, linear, quadratic effects, and interaction, respectively; and X_i and X_j represent the independent variables.

2.4. GC–MS analysis

GC–MS analysis was conducted using a Thermo Trace 1300 ISQ mass chromatograph equipped with a capillary column (HP-5MS, 30 m × 0.25 mm I.D., 0.25 μm film thickness). The oils were methylated and then separated by n-hexane before the analysis. The Mass spectrometry was performed using the electron-impact ionization (EI) mode with 70 eV of energy. The programs performed to recognize the oil composition, which were tested using the following conditions: the initial temperature was maintained at 80 °C for 7 min and then increased to 180 °C at 10 °C/min (maintained for 3 min). It was then increased to 200 °C at 3 °C/min, followed by an increase to 210 °C at 1 °C/min (maintained for 2 min), then an increase to 260 °C at 8 °C/min (maintained for 5 min), and then holding at 260 °C until to the end of the experiment. The injection temperature was 250 °C. Helium was applied as the eluent gas at a 1-mL/min flow speed. For the GC–MS analysis, 1 μL of the sample with a split ratio of 1: 20 was used, and the oil compositions were determined using the National Institute of Standards and Technology spectral library. The amount of fatty acids was calculated using the peak area normalization law and conveyed as the relative proportion of each fatty acid to the total area of all fatty acids.

2.5. Evaluation of the properties of S. Sebiferum seed kernel oil

The physicochemical characteristics of S. sebiferum seed kernel oil were investigated, including the specific gravity, acid, saponification, and iodine values using American Oil Chemists’ Society (AOCS, 2004).

2.6. Statistical analysis

The PBD and BBD were conducted using Design Expert 8.0 (Stat-Ease, Minneapolis, USA), and the results are reported as the mean values. An analysis of variance (ANOVA) test was conducted to assess the significance of the variations in the oil yields. Each experiment was performed in triplicate, and the results are presented as the means ± SD.

3. Results and discussion

3.1. Selection of enzyme

The cell structures of the raw plant materials directly affected the choice of enzyme [19]. Fig. 1 shows that the S. sebiferum seed kernel oil yield changed considerably according to the enzyme type used, with the neutral proteinase yielding the largest amount of oil among all the enzymes. The S. sebiferum seed kernel mainly consists of protein and oil [20]. The oil is released by adding proteases during extraction, directly hydrolyzing and disrupting the protein networks [21], [22], [23]. Therefore, neutral proteinase was selected for optimization of the AEUCE process.

Fig. 1.

Selection of enzyme in the AEUCE process for extracting S. sebiferum seed kernel oil.

3.2. Screening the significant variables with the PBD

The PBD was employed to screen the vital factors that affected the AEUCE procedure (see Table 1). The equation for determining the S. sebiferum seed kernel oil yield (Y) is as follows:

Y = 77.48 + 2.03A + 0.11B + 0.36C + 0.93D + 1.13E + 0.27F (3)

As presented in Fig. 2 and Table 1, the enzyme loading amount (A), ultrasound time (D), and ultrasound power (E) had considerable effects on the oil yield from AEUCE, as all their T values were over the Bonferroni limit (4.8819). The higher enzyme loading amount provides the raw plant materials with more contact with the enzyme, accelerating the release of the S. sebiferum seed kernel oil. However, an increased amount of enzyme may result in unnecessary waste in terms of resources and costs. Herein, the specific needs of the experiment determined the suitable amount of enzyme [24], [25]. Moreover, a longer ultrasound time and higher ultrasound power produce more cavitation bubbles that can achieve optimal growth (black radius), and they then collapse owing to the production of strong micro-jets and shock waves [26], [27]. Therefore, A, D, and E were selected for further optimization in the BBD.

Fig. 2.

PBD for S. sebiferum seed kernel oil extraction by AEUCE. (A) enzyme loading amount (%); (B) incubation time (h); (C) water-to-solid ratio (mL/g); (D) ultrasound time (min); (E) ultrasound power (W); (F) ultrasound temperature (°C).

The incubation time had a statistically insignificant influence on the yield (P > 0.05), which may have been because an appropriate incubation time means that the enzyme makes full contact with the raw plant material. In contrast, a shorter or much longer incubation time will lead to either incomplete extraction or additional products, respectively, reducing the oil yield [28], [29]. Regarding the effect of the water-to-solid ratio, proper solvents promote the enzyme diffusion and enzymatic process. Still, a large amount of solution has a dilution effect on the enzyme and substrate concentrations, reducing the chance of the enzyme and raw plant materials from making contact [30], [31]. Regarding the ultrasound temperature, the solubility and diffusivity of the compounds increased with an increase of ultrasound temperature, enhancing the mass transfer capacity and oil yield. In contrast, higher ultrasound temperatures can reduce the difference in the vapor pressure between the inside and outside of the cavitation bubbles, reducing the intensity of bubble collapse and preventing the complete release of the target compound [32], [33], [34]. Based on the pre-test experiments, the factors that had insignificant impacts were 3 h of incubation time, an ultrasound temperature of 50 °C, and 8 mL/g as the water-to-solid ratio.

3.3. Optimizing the AEUCE process using the BBD

3.3.1. Model fit and statistical analysis

The BBD was used to acquire the optimized conditions for AEUCE. Table 2 details the design matrix for the BBD and the actual and predicted values. A multiple regression analysis was conducted on the oil yield (Y) using the following quadratic polynomial equation:

Y = 80 + 1.78A + 1.00D + 1.36E + 0.86AD + 1.18AE + 0.61DE – 1.04A² – 0.80D² – 0.46E² (4)

Table 2.

Coded and real levels of screened parameters and observed responses.

Runa	A	D	E	Yield (%)
				Actual	Predicted
1	3	30	490	83.18	82.83
2	2	20	350	77.12	77.00
3	2	30	420	79.69	80.00
4	2	30	420	80.73	80.00
5	3	30	350	77.64	77.74
6	2	20	490	78.17	78.50
7	2	30	420	79.88	80.00
8	1	30	490	77.01	76.91
9	1	20	420	76.47	76.25
10	1	30	350	76.21	76.56
11	2	30	420	80.25	80.00
12	2	40	350	78.11	77.78
13	3	20	420	78.06	78.08
14	3	40	420	81.58	81.80
15	2	30	420	79.47	80.00
16	1	40	420	76.55	76.53
17	2	40	490	81.59	81.72

^aA: enzyme loading amount, %; D: ultrasound time, min; E: ultrasound power, W.

Table 3 presents the results of the ANOVA obtained from the BBD. A model F value of 32.64 coupled with a very low P value (P < 0.0001) indicates that the proposed model showed a significant result. The coefficient of determination (R²), adjusted coefficient of determination (Adj. R²), and coefficient of variation (C.V.) were used to assess the model’s goodness-of-fit [35]. The R² was 0.9787, which indicated that the developed model adequately represents the real correlation between the selected variables [36]. The predicted R² of 0.8350 aligned well with the adjusted R² of 0.9468, and the signal-to-noise ratio of 17.968 was much higher than 4, demonstrating that the model exhibited good accuracy. Regarding the influences of linear, interaction, and quadratic parameters on the S. sebiferum seed kernel oil yield in the AEUCE process, the results of the model fitting indicated that all the independent parameters had extremely significant effects, and interactive parameters of AD, AE, and A² had highly significant effects, and the interactive parameter DE and the quadratic parameter D² had significant effects.

Table 3.

Analysis of variance (ANOVA) for the experimental results.

Source	Sum of squares	Degrees of freedom	Mean square	F value	P value Prob > F	Inference
Modela	67.06	9	7.45	32.64	< 0.0001	***
A	25.28	1	25.28	110.70	< 0.0001	***
D	8.02	1	8.02	35.12	0.0006	***
E	14.77	1	14.77	64.69	< 0.0001	***
AD	2.96	1	2.96	12.96	0.0087	**
AE	5.62	1	5.62	24.60	0.0016	**
DE	1.48	1	1.48	6.47	0.0385	*
A2	4.54	1	4.54	19.88	0.0029	**
D2	2.70	1	2.70	11.82	0.0109	*
E2	0.87	1	0.87	3.83	0.0912	ns
Residual	1.60	7	0.23
Lack of fit	0.61	3	0.20	0.83	0.5441	ns
Pure error	0.99	4	0.25
Cor totalb	68.66	16
Std. Dev.c	C.V.d %	Mean	R2	Adjusted R2	Predicted R2	Adequate precision
0.48	78.92	0.61	0.9767	0.9468	0.8350	17.968

^aA: Enzyme loading amount, %; D: ultrasound time, h; E: ultrasound power, W;

^bCor Total: Totals of all information corrected for the mean;

^cStd. Dev.: Standard deviation;

^dC.V.: Coefficient of variation.

3.3.2. Response contour plot

RSM is vital for visualizing the interactions between the independent variables and responses [37]. Fig. 3 details the three-dimensional surface plots established using the predicted second-order polynomial model. As shown on Fig. 3a, the yield of S. sebiferum seed kernel oil improved when the enzyme loading amount and ultrasound time were increased. However, increasing the enzyme loading amount and ultrasound time did not increase the oil yield. This may have been because an appropriate amount of enzyme promotes the enzymatic hydrolysis of substances and stimulates the release of oil from plant materials. However, an excessive enzyme loading amount may lead to extraction imbalance or even a decrease in yield [38], [39]. In addition, long extraction times may adversely affect extraction due to the possible degradation of the oil [40]. As shown in Fig. 3b, the oil yield increased when the enzyme loading amount and ultrasound power were increased. The highest yield (82.83 %) was obtained with an enzyme loading amount of 2.92 % and an ultrasound power of 490 W. Fig. 3c shows that the oil yield first increased with an increase in ultrasound time and power. An additional increase in these two parameters did not increase the yield. A maximum oil yield (84 %) was acquired when the ultrasound time and power reached 38.59 min and 490 W, respectively. This is consistent with previous studies, which may be due to the oxidation of target compounds and the loss caused by high ultrasound power [41].

Fig. 3.

Three-dimensional surfaces generated from the BBD. (a) The interactive effects of the enzyme loading amount (A) and ultrasound time (D) on the S. sebiferum seed kernel oil yield; (b) the interactive effects of the enzyme loading amount (A) and ultrasound power (E) on the S. sebiferum seed kernel oil yield; and (c) the interactive effects of ultrasound time (D) and power (E) on the S. sebiferum seed kernel oil yield.

3.3.3. Method validation

The BBD results indicate that the optimal AEUCE conditions for a maximum S. sebiferum seed kernel oil yield of 83.33 % are 2.95 % enzyme loading amount, 38.96 min of ultrasound time, and 485 W of ultrasound power. Considering the operability, the conditions for the AEUCE process were slightly modified to an ultrasound time of 39 min. Triplicate experiments were then conducted to confirm the precision of the regression model, from which an oil yield of 84.22 ± 3.17 % was obtained, which aligns with the BBD’s predicted S. sebiferum seed kernel oil yield. These results indicate that the proposed model is feasible and accurate for optimizing the AEUCE process to separate S. sebiferum seed kernel oil.

AEUCE was first proposed to extract the seed kernel oil of S. sebiferum, which performed better than that of SE in terms of time and the organic solvents employed. This is likely because of two reasons. 1) In the AEUCE process, the cell walls of plant materials are ruptured by the added enzyme, which could increase the cell permeability and facilitate the release of oil [42]. 2) The effects of the vibrations and cavitation caused by the ultrasound waves could produce many cavitation bubbles, which gradually gather and burst on the surface of the material with high enough amplitudes to rupture the cell walls, thus accelerating the transfer of the desired compounds from the plant cells to the solvent [16].

3.4. Characterization of S. Sebiferum seed kernel oil

3.4.1. Fatty acid composition

The chemical compositions of the S. sebiferum seed kernel oils extracted via AEUCE and SE are shown in Table 4. Ten constituents were identified in the S. sebiferum seed kernel oil, including saturated fatty acids (SFAs), monounsaturated fatty acids (MUFAs), and polyunsaturated fatty acids (PUFAs). The proportions of these three fatty acids in the oil obtained from AEUCE were very similar to those obtained in SE. The most abundant component of the SFAs in the seed oil was palmitic acid (C_16:0; 5.67 % for AEUCE and 5.89 % for SE). The others were stearic acid (C_18:0), decanoic acid (C_10:0), lauric acid (C_12:0), and myristic acid (C_14:0). Among the MUFAs, oleic acid (C_18:1,n–9) and eicosenoic acid (C_20:1,n–9) were the most abundant components in the S. sebiferum seed kernel oil. Finally, linoleic acid (C_18:2, 23.66 % for AEUCE and 25.01 % for SE) and linolenic acid (C_18:3, 64.33 % for AEUCE and 62.23 % for SE) were the main PUFAs. Research indicates that the AEUCE process generates a powerful shockwave and a high-speed jet that ruptures cell walls, allowing the release of more oil [43]. Wang et al. determined the different profiles of fatty acid components in other S. sebiferum seed kernel oils, in which lauric acid (C_12:0) and decanoic acid (C_10:0) were not detected [44]. These differences in the fatty acid profiles of the S. sebiferum seed kernel oil may be due to the extraction methods or the environmental conditions in which S. sebiferum grew. Moreover, the high unsaturated fatty acid levels in S. sebiferum seed kernel oil make it appropriate for biodiesel production as it ensures proper cold flow peculiarity (e.g., pour point and cold filter plugging point) [45], [46]. These results indicate that AEUCE is a promising alternative to traditional SE for separation of S. sebiferum seed kernel oil. The oil obtained from this method has great promise for biodiesel production.

Table 4.

Chemical composition of seed kernel oil from S. sebiferum by GC–MS analysis.

No.	Compounds	Retention time (min)	Molecular formula	Relative peak area (%)
				AEUCE	SE
1	Decanoic acid (C10:0)	13.69	C10H20O2	0.04	0.11
2	2,4-Decadienoic acid (C10:2)	14.81	C10H16O2	2.32	2.83
3	Lauric acid (C12:0)	16.58	C12H24O2	0.01	0.08
4	Myristic acid (C14:0)	19.56	C14H28O2	0.04	0.03
5	Palmitic acid (C16:0)	24.10	C16H32O2	5.67	5.89
6	Linoleic acid (C18:2)	28.66	C18H32O2	23.66	25.01
7	Linolenic acid (C18:3)	28.89	C18H30O2	64.33	62.23
8	Oleic acid (C18:1)	29.03	C18H34O2	2.08	1.88
9	Stearic acid (C18:0)	29.72	C18H36O2	1.69	1.77
10	Eicosenoic acid (C20:1)	36.78	C20H38O2	0.16	0.17
Total (%)				100	100
Saturated fatty acid (%)				7.45	7.88
Monounsaturated fatty acid (%)				2.24	2.05
Polyunsaturated fatty acid (%)				90.31	90.07
Degree of unsaturation				182.86	182.19

3.4.2. Physical and chemical properties

Specific gravity is an important factor for evaluating the purity of plant oils. Table 5 shows these results, indicating that the oils obtained by AEUCE and SE were of similar purities. Furthermore, the acid value is an important indicator of biodiesel feedstock, where a low acid value is beneficial for shelf life and the transesterification of biodiesel [46], [47]. Table 5 demonstrates that SE yielded an oil with a significantly higher acid value (4.90 ± 0.10 mg KOH/g) than AEUCE (4.61 ± 0.08 mg KOH/g), illustrating that the S. sebiferum seed kernel oil extracted by AEUCE is an ideal raw oil for the production of biodiesel compared to that obtained from SE. The iodine value signifies the unsaturation levels and the chain length of the fatty acid in the triacylglycerols [48], [49]. No significant differences were found between the iodine values of the oils obtained from AEUCE and SE, indicating that the two oils had similar unsaturated fatty acid levels in triacylglycerols.

Table 5.

Physicochemical properties of seed kernels oils from S. sebiferum extracted by AEUCE and SE.

	AEUCE	SE
Specific gravity (15 °C, g/mL)	0.87 ± 0.02a	0.89 ± 0.01a
Acid value (mg KOH/g)	4.61 ± 0.08b	4.90 ± 0.10a
Saponification value (mg/g)	204.93 ± 2.17a	206.25 ± 2.81a
Iodine value (g I2/100 g)	186.85 ± 3.58a	182.97 ± 3.72a

Each test was performed in three times and values are means ± SD. ANOVA was employed for data analysis and values annotated with the letter are not significant at the 5 % level (P < 0.05).

3.5. SEM morphologies

SEM was employed to determine the morphologies of the S. sebiferum seed kernels before and after extraction, as described previously [50]. The SEM images of S. sebiferum seed kernels that were untreated and extracted by AEUCE and SE are presented in Fig. 4. Fig. 4a shows that the cell walls of the raw plant materials were intact, as indicated by the smooth surface. After they were extracted by the AEUCE process (Fig. 4c), the sample cells were considerably destroyed. They had some irregular holes on their surfaces due to the cavitation effect. In contrast, these changes were not observed on the samples extracted by SE (Fig. 4b). This is likely due to the synergetic effects of the enzymatic hydrolysis and cavitation caused by the ultrasound waves [51], [52]. A similar phenomenon was found in the ultrasound extraction process for mahua seed oil and Cinnamomum camphora seed oil [16], [50].

Fig. 4.

SEM images of the untreated S. sebiferum seed kernels (a) and those extracted by SE (b) and AEUCE (c).

4. Conclusion

S. sebiferum seed kernel oil is nonedible and rich in unsaturated fatty acids, indicating that it could be an ideal feedstock oil for biodiesel production. In this study, we used AEUCE to separate the oil from S. sebiferum seed kernels, and neutral proteinase was employed as the enzyme. Optimal AEUCE conditions were attained using the PBD and BBD, which resulted in a yield of 84.22 ± 3.17 % for the following conditions: 2.95 % enzyme loading amount, 3 h of incubation time, 8 mL/g water-to-solid ratio, 39 min of ultrasound time, 485 W of ultrasound power, and an ultrasound temperature of 50 °C. Apart from their content, no discernible differences were observed in the chemical composition of the oils extracted by AEUCE. The S. sebiferum kernel seed oil extracted by AEUCE had a lower acid value and a slightly higher degree of unsaturation than that from SE. The morphological structure changes determined by SEM indicated that AEUCE could rupture the seed kernel’s cell walls, accelerating the separation of the oil. Based on these results, AEUCE is an eco-friendly approach that could be used to separate high-quality S. sebiferum seed kernel oil for biodiesel production.

CRediT authorship contribution statement

Zaizhi Liu: Conceptualization, Writing – review & editing, Data curation, Funding acquisition, Supervision. Haibin Liao: Investigation, Methodology, Data curation. Cheng Wei: Investigation, Methodology. Yanlong Qi: Software, Methodology, Investigation, Visualization. Zhengrong Zou: Validation, Funding acquisition, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.