Effects of different extraction methods on the trace components and volatile composition, antioxidant activity of Idesia polycarpa fruit oil

Abstract

This study evaluated the effects of five extraction methods (fresh pressing, cold pressing, hot pressing, solvent extraction, and aqueous extraction) on the quality of Idesia polycarpa oil (IPO). The results revealed that the solvent–extracted oil (SEO) exhibited the highest efficiency. However, the fresh–pressed oil (FPO) exhibited outstanding quality characteristics. Specifically, the acid value and peroxide value of FPO were the lowest, and the retention rate of beneficial trace components such as α–tocopherol, squalene, β–sitosterol, stigmasterol, and carotenoids were rather prominent. It is worth noting that the five IPOs were similar in fatty acid composition and content (all rich in unsaturated fatty acids mainly composed of linoleic acid). Through analysis of volatile components and Orthogonal Partial Least Squares–Discriminant, 14 key biomarkers were identified. There were significant differences in the antioxidant capacity of five IPOs, and the aqueous–extracted oil (AEO) stands out the most, indicating that AEO is more effective at resisting radicals produced during oxidation. In conclusion, this study offers valuable data references for selecting extraction processes rationally in the oil industry.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s10068-025-01994-w.

Introduction

Idesia polycarpa, a woody oil plant belonging to the family Flacourtiaceae, is extensively distributed across middle–to–high–altitude regions in Shaanxi, Gansu, and Henan provinces of China (Ghorbanzadeh and Rezaei, 2017). I. polycarpa is renowned for its high oil content in fruits (Zuo et al., 2024). This characteristic makes it a highly promising woody oil crop. Study has demonstrated that the oil content of dried I. polycarpa fruits ranges from 20 to 45%, with its fatty acid composition primarily consisting of linoleic acid, oleic acid, and palmitic acid (Luo et al., 2025). Notably, linoleic acid, a polyunsaturated fatty acid, constitutes up to 70% of the total fatty acids (Xiang et al., 2023). Research has confirmed that linoleic acid plays a critical role in reducing low–density lipoprotein (LDL) cholesterol levels and preventing cardiovascular diseases (Luo et al., 2025). Furthermore, I. polycarpa oil (IPO) is abundant in bioactive compounds such as squalene, vitamin E, and sterols, which contribute to its potential health benefits, including delaying aging and enhancing blood circulation (Huang et al., 2023). This indicates that the IPO has great development potential and market prospects.

Extraction methods plays a crucial role in determining the quality of oil. Previous studies have demonstrated that different extraction methods significantly influence the micronutrients, nutritional quality, volatile compounds, and lipid profiles of oils (Li et al., 2024). Pressing is one of the most commonly used extraction techniques, characterized by its simplicity, high degree of automation, and suitability for large–scale production. Depending on the pressing temperature, pressing can be categorized into cold pressing and hot pressing. These two methods differ in terms of oil yield, physicochemical properties, and the retention of bioactive components. Cold pressing is more effective at preserving heat–sensitive components such as tocopherols in walnut oil (Xu et al., 2023). Additionally, the profiles of major volatile compounds in cold–pressed walnut oil differ significantly from those in hot–pressed walnut oil (Xu et al., 2023). The peroxide value of hot–pressed safflower oil (at 160 °C) is approximately five times higher than that of cold–pressed safflower oil (at 25 °C), while hot–pressed safflower oil contains significantly higher levels of tocopherols and sterols compared to cold–pressed safflower oil (Hou et al., 2024). Although traditional pressing is widely applied in the oil industry, it imposes strict requirements on the moisture content of plant materials (about 7–10%) (Martinez et al., 2013). The fresh pressing method, characterized by direct extraction of undried fruits, effectively retards lipid oxidation and hydrolytic degradation. This method is standard practice in olive oil production (Derardja et al., 2025). Aqueous extraction, a non–mechanical traditional method, is commonly employed for sesame oil production (Huang et al., 2024). This method relies on high–temperature roasting of materials followed by hot water extraction. Solvent extraction leverages the “like dissolves like” principle to provide mild processing conditions, high efficiency, and enhanced retention of oil nutrients. These advantages drive its growing industrial adoption. (Zhang et al., 2023a).

Currently, research on the extraction of IPO predominantly focuses on single methods (Luo et al., 2025). However, the impact of extraction methods on the holistic quality of IPO remains unelucidated, and a definitive evaluation of its optimal extraction process is lacking. Based on the above considerations, we conducted the following work: (1) Five extraction methods (fresh pressing, cold pressing, hot pressing, solvent extraction, and aqueous extraction) were employed to obtain IPOs. (2) The physicochemical properties of IPOs were measured and analyzed. (3) Gas chromatography was used to analyze the composition of squalene, sterols, and fatty acids. High–performance liquid chromatography (HPLC) and colorimetry were utilized to analyze multiple lipid concomitants in IPOs. Headspace solid–phase microextraction coupled with gas chromatography–mass spectrometry (HS–SPME–GC–MS) technology was applied to investigate the volatile components in IPOs. (4) The chemical structure, rheological property and antioxidant capacity of IPOs were determined. This study aims to elucidate the effects of different extraction methods on the quality of IPOs and provide scientific data to support the industrialization of IPO.

Materials and methods

Materials and chemicals

The fresh fruits of I. polycarpa were supplied by the Shanlinyuan Forestry Comprehensive Development Co., Ld. (Chongqing (28.93 N, 106.97 E), China). Solvent–extracted oil No. 6 (It is mainly composed of n–hexane and cyclohexane) was procured from Dongguan Shizhengxin Chemical Technology Co., Ltd. (Guangdong, China). Rutin, gallic acid and Folin–Ciocalteu were provided by Solarbio Science & Technology Co., Ltd. (Beijing, China). Squalene, 2–octanol, α–tocopherol, β–sitosterol, stigmasterol, DPPH, ABTS and fatty acid standard were purchased from Macklin Biochemical Technology Co., Ltd. (Shanghai, China). Methanol, isopropanol, n–hexane, ethanol, petroleum ether, and all other chemical reagents used in this study were purchased from Aladdin Biochemical Technology Co., Ltd. (Shanghai, China).

Pre–treatment of I. polycarpa fruits

Fresh I. polycarpa fruits were washed and air–dried. One portion was stored at –80 °C, while another portion was oven–dried at 50 °C, powdered (RS–FS1411, Royalstar, Hefei, China) and sieved through a 40–mesh screen for subsequent research.

IPO extraction

The steps for extracting IPO using five different methods are presented in Fig. 1. For fresh pressing, fresh fruits were homogenized into pulp using a grinder (JYL–523, Joyoung, China). For cold and hot pressing, an oil extraction machine (D02, Han Huang, China) was preheated to 50 and 150 °C, respectively. Additionally, dried fruits were fried for 10 min prior to hot pressing. The solvent extraction experiment was conducted following the method of Zhang et al. (2023b). Rotary evaporation (V–700, Buchi, Switzerland) was performed under predetermined conditions (80 mbar, 40 °C). Aqueous extraction was carried out as described by Huang et al. (2024). All extraction methods were centrifuged (4226×g, 10 min) to isolate the upper oil phase, yielding: fresh–pressed oil (FPO), cold–pressed oil (CPO), hot–pressed oil (HPO), solvent–extracted oil (SEO), and aqueous–extracted oil (AEO). All IPOs were stored at 4 °C to prevent oxidation.

Fig. 1.

Flowchart of five methods for extracting IPOs

Determination of extraction rate

The extraction rate of each method can be calculated using the following formula (1):

$$Extractionrate(\text{\%})=\frac{m_1}{m_0}\times 100$$ 1

where m₁ refers to the mass of the IPO extracted by each method, and m₀ represents the mass of the IPO obtained by extracting the same mass of materials using Soxhlet extraction.

Determination of physicochemical properties

The color parameter (L^*, a^*, b^*) of the IPOs were measured using a colorimeter (CM–5, Konica Minolta, Japan). The acid value, peroxide value, saponification value, and unsaponifiable matter of IPO were determined following the method described by Liang et al. (2025).

Determination of trace components

It is worth noting that the content of trace components in the IPOs was qualitatively and quantitatively analyzed by the external standard method.

Total polyphenols and total flavonoids

The determination of total polyphenols and total flavonoids was based on the method of Likhar et al. (2025). For total polyphenols content: 0.5 g IPO was mixed with 2.5 mL Folin–Ciocalteu reagent, incubated in darkness (30 min), and absorbance measured at 765 nm by Microplate Reader (011005, BioTek., USA). For flavonoid quantification: 0.1 g IPO was sequentially reacted with 1 mL NaNO₂, 1 mL Al (NO₃)₃, and 10 mL NaOH under dark conditions (30 min), with subsequent absorbance measurement at 510 nm.

α–Tocopherol

The content of α–tocopherol was determined according to the method described by Xiang et al. (2023). IPO (1 g) was dissolved in absolute ethanol (5 mL), thoroughly mixed, and subjected to ultrasonication for 20 min. The solution was then filtered through a 0.22 μm organic filter membrane. Subsequently, 20 μL of the filtrate was injected into an HPLC system (1260, Agilent, USA) equipped with an Agilent Zorbax C18 column (250 mm × 4.6 mm, 5 μm) and a UV detector (λ = 294 nm). The column temperature was maintained at 30 °C, and n–hexane/methanol/isopropanol (92.5/7.4/0.1, v/v/v) was used as the mobile phase for elution at a flow rate of 1 mL/min.

Squalene, β–sitosterol and stigmasterol

The contents of squalene and sterols were determined in accordance with Majcher et al. (2025). 0.2 g of IPO was weighed and mixed with 1 mL of KOH–ethanol solution, 5 mL of n–hexane, and 10 mL of saturated NaCl solution. Following extraction, 3 mL of the n–hexane phase was passed through a 0.22 μm organic filter membrane. Subsequently, 1 μL of the filtrate was injected into a GC system (7080A, Agilent, USA) equipped with a DB–5 capillary column (0.25 μm, 30.0 m × 0.35 mm) and a flame ionization detector (FID). The nitrogen carrier gas flow rate was 1 mL/min. The temperatures of the injector and detector were set at 250 and 300 °C, respectively. The GC heating procedure was as follows: initial temperature 160 °C held for 1 min; ramped to 220 at 15 °C/min and held for 2 min; then ramped to 280 at 5 °C/min and held for 10 min; finally ramped to 300 at 10 °C/min and held for 2 min.

Carotenoids and chlorophyll

Referring to the method described by Fan et al. (2023). 300 mg of IPO was mixed with 10 mL of n–hexane and vortex–mixed for 1 min, followed by ultrasonic extraction at 40 °C for 20 min. The absorbance of the mixture was measured at 470 and 670 nm, respectively. The contents of carotenoids and chlorophyll in the extract were calculated using formulas (2) and (3):

$$Carotenoids(\text{mg/100g})=(A_{470}\times {10}^6)/(1000\times 2000\times C)$$ 2

$$Chlorophyll(\text{mg/100g})=(A_{670}\times {10}^6)/(1000\times 613\times C)$$ 3

where C denotes the concentration of the IPO (g/mL).

Determination of fatty acid composition

The IPOs were saponified and methylated following by Zhang et al. (2023a). IPO (0.5 g), pyrogallic acid (100 mg), 95% ethanol (2 mL), and NaOH–ethanol solution (15 mL) were sequentially added to the flask. The mixture was thoroughly mixed and refluxed at 80 °C until oil droplets disappeared. Subsequently, 15% boron trifluoride–methanol solution (7 mL) was added, followed by distilled water (5 mL), after which the flask was removed. Finally, n–heptane (20 mL) and saturated NaCl solution (10 mL) were introduced, and the upper organic layer was transferred to an injection vial.

Analysis was conducted using an SP–2560 column (100 m × 0.25 mm × 0.2 μm) with nitrogen carrier gas (1 mL/min). The injector and FID temperatures were set to 220 and 280 °C, respectively. The heating procedure was as follows: 100 °C (held for 13 min), increased at 10 °C/min to 180 °C (held for 6 min), then at 1 °C/min to 200 °C (held for 20 min), and finally at 4 °C/min to 230 °C (held for 10 min). Single fatty acid was determined by comparing the retention times of IPOs with fatty acid standard, and the relative content (%) of each fatty acid was computed using the area normalization method.

Determination of volatile compounds

Volatile compounds were determined according to the method described by Zhang et al. (2025a). Two grams of IPO were weighed and placed in a 20 mL headspace vial, 20 μL of 2–octanol solution (10 μg/μL) was added as the internal standard. The vial was heated at 60 °C for 15 min, followed by a 30–min extraction period. The headspace was subsequently injected into the GC system (250 °C) equipped with a DB–5MS capillary column (30 m × 0.25 mm, 0.25 µm). The heating procedure was: 60 °C (held 10 min), ramp at 2 °C/min to 70 °C, then 10 °C/min to 140 °C, 5 °C/min to 170 °C (held 1 min), and finally 15 °C/min to 230 °C (held 4 min). Helium was used as the carrier gas at a constant flow rate of 0.8 mL/min (37.8 kPa). MS conditions included electron impact ionization (EI) at 80 eV, ion source and interface temperatures at 230 °C, a scan speed of 769 u/s, and mass scan range of 40–400 (m/z). The GC–MS spectra were compared against the National Institute of Standards and Technology (NIST) database for compound identification. The concentration (μg/g) of each volatile compound was quantified using the internal standard method.

Determination of infrared spectrometric characteristics

Briefly, 20 µL of IPO was analyzed using an infrared spectrometer (Spectrum 100, PerkinElmer, USA) with a spectral range of 3500 to 500 cm⁻¹, employing 32 scans at a resolution of 1 cm⁻¹.

Determination of rheological property

The viscosity of the IPO was determined using the method described by Zhang et al. (2025b), which involved a rheometer (MCR 302, Anton Paar, Austria) equipped with a CC27 rotor. The experimental setup included a plate gap of 1 mm, a shear rate range of 0.1–300 s^–1, and viscosity curve scanning at 25 °C, with 50 data points collected.

Determination of antioxidant activity

DPPH radical scavenging rate

The scavenging capacity of DPPH radical was determined following the method described by Kalkan and Maskan (2025). Briefly, the reaction involved incubating the DPPH standard solution with IPO in the dark for 30 min. The absorbance of the mixture was measured at 517 nm, and the DPPH radical scavenging rate (%) was calculated using formula (4):

$$DPPHradicalscavengingrate(\%)=(1-\frac{A_1-A_2}{A_0})\times 100$$ 4

where A₀ denotes the absorbance of the blank sample (DPPH + ethanol), A₁ indicates the absorbance of the experimental sample (DPPH + IPO), while A₂ represents the absorbance of the control group (IPO + ethanol).

ABTS radical scavenging rate

The scavenging rate of ABTS radical was determined in accordance with the protocol given by Czett et al. (2025). Briefly, the ABTS standard solution and IPO were reacted in the dark for 10 min, and the absorbance of the mixture was measured at 734 nm. The scavenging rate of ABTS radicals was calculated using the formula (5):

$$\text{ABTS radical scavenging rate}(\%)=(1-\frac{{\text{A}}_1-{\text{A}}_2}{A_0})\times 100$$ 5

where A₀ denotes the absorbance of the blank sample (ABTS + ethanol), A₁ indicates the absorbance of the experimental sample (ABTS + IPO), and A₂ indicates the absorbance of the control group (IPO + ethanol).

Total reducing power

The total reducing power was determined using the method described by Feng et al. (2016). Briefly, IPO, acid buffer (pH = 6.6), and potassium ferricyanide were added to a test tube successively and reacted at 50 °C for 20 min, followed by the addition of trichloroacetic acid solution. The supernatant was collected, and deionized water and FeCl₃·6H₂O solution were added successively. The absorbance of the mixture was measured at 700 nm.

Statistical analysis

All experiments were conducted in triplicate, and reported as the mean ± standard deviation (SD). Statistical analysis was performed using SPSS Statistics 26.0 software (IBM SPSS, Chicago, USA). Significance was determined by one–way analysis of variance (ANOVA), and differences among means were analyzed using Duncan’s multiple range test (p < 0.05).

Results and discussion

Physicochemical properties

As shown in Table 1, extraction rate varied significantly among methods (39.23–84.02%). Solvent extraction achieved the highest rate (84.02%), while fresh pressing yielded the lowest (39.23%). The superior efficiency of solvent extraction for IPO likely stems from the close dielectric constants of nonpolar solvents and triglycerides, facilitating accelerated oil dissolution and enhanced extraction rate (Zhang et al., 2019). The extraction rate of aqueous extraction (40.68%) was found to be comparable to that of fresh pressing extraction. Both methods demonstrated relatively low extraction efficiency, attributable to protein adsorption at the oil–water interface during extraction. This phenomenon formed a stable protein layer that impeded IPO coalescence, thereby hindering IPO collection and reducing the extraction rate (Wang et al., 2023). The extraction rate of hot pressing (61.15%) significantly exceeded that of cold pressing (54.10%). This disparity was attributed to high–temperature–induced softening of plant cell wall cellulose and disruption of lipoprotein complexes, which liberated encapsulated IPO (Wu et al., 2024). The L^* values of the IPOs displayed notable differences, with CPO being the brightest and AEO the dimmest (Table 1). This might be because during high-temperature water extraction, more pigments from the powder of I. polycarpa become concentrated in the AEO, thereby deepening its color (lower L^* value) (Huang et al., 2024). The a^* values suggested that CPO and HPO had a darker reddish–brown, while the b^* values indicated a deeper yellowness in CPO and HPO compared to the relatively lighter yellowness of AEO. These color discrepancies might be attributed to variations in pigment content and other trace components. Notably, FPO exhibited consistently low acid value (1.44 mg KOH/g) and peroxide value (3.84 mmol/kg), indicating minimal formation of free fatty acids and peroxides during extraction, thereby meeting the standards for edible oils. Additionally, this likely reflected the absence of heat–drying pretreatment in fresh fruits, which consequently exhibited reduced oxidation. This observation further indicated that both cold pressing and hot pressing extraction promoted primary lipid oxidation, generating substantial hydroperoxides that directly elevated peroxide values in CPO and HPO.

Table 1.

Extraction rate and physicochemical properties of IPOs extracted by different methods

Physicochemical index	FPO	CPO	HPO	SEO	AEO
Extraction rate (%)	39.23 ± 0.02d	54.10 ± 0.05c	61.15 ± 0.05b	84.02 ± 0.01a	40.68 ± 0.10d
L*	52.84 ± 0.06c	57.43 ± 0.01a	56.13 ± 0.01b	48.26 ± 0.01d	45.01 ± 0.09e
a*	48.07 ± 0.04c	52.60 ± 0.02a	52.63 ± 0.01a	47.31 ± 0.01d	49.99 ± 0.03b
b*	91.10 ± 0.10c	99.02 ± 0.02a	96.77 ± 0.02b	83.21 ± 0.02d	77.60 ± 0.15e
Acid value (mg KOH/g)	1.44 ± 0.62c	8.48 ± 1.61b	9.10 ± 0.38b	11.92 ± 1.14a	12.91 ± 0.50a
Peroxide value (mmol/kg)	3.87 ± 0.04c	21.82 ± 1.28a	22.16 ± 1.73a	9.98 ± 1.26b	7.9 ± 1.07b
Saponification value (mg KOH/g)	201.69 ± 2.15b	196.49 ± 2.19bc	191.57 ± 2.49c	155.34 ± 1.10d	216.62 ± 2.29a
Unsaponifiable matter (%)	0.78 ± 0.11a	0.63 ± 0.08ab	0.46 ± 0.02bc	0.37 ± 0.04c	0.72 ± 0.11a

Different letters (a–e) in the same row indicate statistically significant differences (p < 0.05)

FPO fresh–pressed oil, CPO cold–pressed oil, HPO hot–pressed oil, SEO solvent–extracted oil, AEO aqueous–extracted oil

There were considerable differences in the saponification values among the IPOs, with AEO exhibiting the highest value (216.62 mg KOH/g) and SEO the lowest (155.34 mg KOH/g). AEO exhibited elevated acid and saponification values. This was attributed to high–temperature aqueous conditions promoting triglyceride hydrolysis, thereby exposing additional saponifiable ester bonds. The unsaponifiable matter content of IPOs extracted by different processes ranges from 0.37 to 0.78%, with FPO having the highest value. This may be attributed to the higher concentration of bioactive components in fresh fruits.

Trace components

As shown in Table 2, the extraction methods significantly influenced the total polyphenol and flavonoid contents of the IPOs. Among the five IPOs, total polyphenol content followed the order: AEO > HPO > CPO > SEO > FPO. HPO exhibited the highest total flavonoid content, followed by CPO and SEO. Previous studies indicated that hot pressing not only increases oil yield but also promotes the migration of bioactive compounds into the oil (Fan et al., 2023). The higher levels of both total polyphenol and flavonoid content in HPO compared to CPO can likely be attributed to high–temperature disruption of plant cell integrity, which enhances the solubility of these compounds in the oil.

Table 2.

Bioactive components content of five IPOs

Parameters	FPO	CPO	HPO	SEO	AEO
Total polyphenols (mg GAE/100 g)	45.68 ± 0.33e	53.36 ± 0.34c	79.09 ± 1.47b	49.88 ± 1.32d	90.68 ± 2.52a
Total flavonoids (mg RE/100 g)	4.79 ± 0.26d	13.27 ± 0.53b	15.34 ± 1.27a	12.05 ± 0.77bc	11.27 ± 0.30c
α–Tocopherol (mg/100 g)	43.65 ± 0.21a	15.59 ± 0.20c	16.12 ± 0.06c	17.20 ± 0.69b	18.02 ± 0.37b
Squalene (mg/100 g)	22.64 ± 0.62a	21.10 ± 3.06ab	20.87 ± 2.95ab	16.27 ± 0.34b	20.23 ± 1.37ab
β–sitosterol (mg/100 g)	290.38 ± 2.93a	272.27 ± 8.17b	283.86 ± 2.02ab	239.72 ± 8.24c	281.27 ± 9.60ab
Stigmasterol (mg/100 g)	7.49 ± 0.16a	2.61 ± 0.09c	1.1 ± 0.17e	3.18 ± 0.08b	2.25 ± 0.15d
Carotenoids (mg/100 g)	26.57 ± 0.50a	19.98 ± 1.63d	22.01 ± 0.55b	12.37 ± 1.25e	20.80 ± 0.29c
Chlorophyll (mg/100 g)	4.50 ± 0.77b	4.41 ± 0.74b	4.00 ± 0.09b	5.77 ± 0.47a	4.39 ± 0.21b

Different letters (a–e) in a row indicate significant differences (p < 0.05)

FPO fresh–pressed oil, CPO cold–pressed oil, HPO hot–pressed oil, SEO solvent–extracted oil, AEO aqueous–extracted oil

In the IPO, the predominant form is α–tocopherol (Xiang et al., 2023). FPO exhibited the highest α–tocopherol content (43.65 mg/100 g), while the other four IPOs showed minimal variation. This was attributed to both the higher initial α–tocopherol content in fresh fruits of I. polycarpa compared to dried fruits, and the mild extraction conditions that reduced tocopherol isomerization.

Similar to the findings of Wu et al. (2025), our research indicated that different extraction methods did not significantly affect the squalene content in IPOs. The squalene content (16.27–22.64 mg/100 g) in IPOs was significantly higher than that in several specialty woody oils, including walnut oil (Pokkanta et al., 2019), palm oil (Rébufa et al., 2022).

Additionally, phytosterols, a type of triterpene compound often regarded as nutritional supplements in healthy oils, possess cholesterol–lowering, anti–inflammatory, and immune–regulatory effects. There were significant differences in the contents of β–sitosterol and stigmasterol among the IPOs, and the sterols were dominated by β–sitosterol. Interestingly, FPO retained the highest content of β–sitosterol (290.38 mg/100 g) and stigmasterol (7.49 mg/100 g). This suggested that the fresh pressing extraction better preserved bioactive components including α–tocopherols, squalene, β–sitosterol.

Among the five IPOs, carotenoid content differed significantly (p < 0.05), with the following descending order: FPO > HPO > AEO > CPO > SEO. This is possibly due to the fact that the fresh fruits of I. polycarpa were not exposed to high–temperature roasting, protecting the double–bond structure of carotenoids and subsequently resulting in their enrichment in FPO during the fresh pressing extraction. Chlorophyll content was notably higher in SEO than in other IPOs (5.77 mg/100 g), might suggesting that solvent extraction favors chlorophyll retention.

Fatty acids composition

The nutritional value, processing characteristics, stability, and application scope of vegetable oils are strongly influenced by their fatty acid composition. As shown in Table 3, the IPOs comprised 12 fatty acids. Linoleic acid (63.36–66.36%) was predominant, followed by palmitic acid (17.81–18.90%) as the second most abundant component. While the extraction methods did not alter the overall fatty acid composition, they resulted in quantifiable differences in the content of specific fatty acids among the five IPOs. FPO exhibited the highest saturated fatty acid level (23.43%) and the lowest unsaturated fatty acid level (76.57%). In contrast, CPO contained the highest proportion of unsaturated fatty acids (77.32%), closely followed by SEO (77.16%). This variation suggests that short-term extraction and milder extraction conditions, as used in CPO and SEO, may better preserve unsaturated fatty acids from oxidation.

Table 3.

Fatty acid composition of five IPOs

No	Fatty acids	Content (%)
		FPO	CPO	HPO	SEO	AEO
1	C14:0	0.12 ± 0.03a	0.10 ± 0.04a	0.09 ± 0.28a	0.09 ± 0.02a	0.07 ± 0.01a
2	C16:0	18.90 ± 0.04a	17.81 ± 0.13b	17.88 ± 0.06b	17.85 ± 0.51b	17.94 ± 0.16b
3	C16:1	4.92 ± 0.05a	4.68 ± 0.01b	4.73 ± 0.04b	4.63 ± 0.06b	4.68 ± 0.09b
4	C17:0	0.28 ± 0.08a	0.20 ± 0.01a	0.27 ± 0.00a	0.21 ± 0.01a	0.24 ± 0.04a
5	C17:1	0.13 ± 0.01a	0.34 ± 0.11b	0.39 ± 0.06b	0.40 ± 0.04b	0.30 ± 0.01b
6	C18:0	1.46 ± 0.03c	1.78 ± 0.01ab	1.78 ± 0.06ab	1.72 ± 0.01b	1.86 ± 0.04a
7	C18:1	6.39 ± 0.13a	4.83 ± 0.02b	4.80 ± 0.06b	4.76 ± 0.04b	4.87 ± 0.08b
8	C18:2	63.36 ± 0.24b	66.06 ± 0.09a	65.60 ± 0.36a	66.36 ± 0.46a	65.65 ± 0.37a
9	C18:3n3	0.78 ± 0.06b	0.98 ± 0.01a	0.98 ± 0.03a	1.03 ± 0.04a	1.02 ± 0.00a
10	C20:0	2.57 ± 0.08a	2.56 ± 0.04a	2.58 ± 0.01a	2.42 ± 0.01b	2.63 ± 0.06a
11	C22:0	0.11 ± 0.01c	0.25 ± 0.01b	0.37 ± 0.06a	0.24 ± 0.01b	0.27 ± 0.01b
12	C22:1	1.02 ± 0.09a	0.46 ± 0.50c	0.39 ± 0.50c	0.32 ± 0.06c	0.68 ± 0.50b
SFAs		23.43 ± 0.01a	22.68 ± 0.04b	22.96 ± 0.09ab	22.52 ± 0.49b	23.00 ± 0.14ab
MUFAs		12.45 ± 0.16a	10.30 ± 0.04b	10.48 ± 0.23b	9.78 ± 0.46b	10.33 ± 0.23b
PUFAs		64.13 ± 0.18b	67.03 ± 0.08a	66.58 ± 0.33a	67.39 ± 0.50a	66.67 ± 0.37a
UFAs		76.57 ± 0.01c	77.32 ± 0.04a	77.05 ± 0.09b	77.16 ± 0.04ab	77.00 ± 0.14b

Different letters (a–c) in a row indicate statistically significant differences (p < 0.05)

FPO fresh–pressed oil, CPO cold–pressed oil, HPO hot–pressed oil, SEO solvent–extracted oil, AEO aqueous–extracted oil

Volatile compounds

Volatile organic compounds (VOCs) in five IPOs were analyzed using GC–MS. Following screening, 62 VOCs were identified and categorized as follows: 6 acids, 14 ketones, 7 alcohols, 9 esters, 4 alkanes, 5 phenols, 12 aldehydes, and 5 furans (Table S1). Among them, the aldehyde and ketone compounds typically originate from the oxidative degradation of unsaturated fatty acids, endowing nuts and baked goods with their pleasant aroma (Xu et al., 2024). For a more intuitive understanding of VOCs differences among IPOs, the compounds were normalized and used to create a clustering heatmap (Fig. 2). We observed that CPO and HPO showed similar clustering trends, which indicated that the VOCs compositions of CPO and HPO were very similar. Table S1 indicated that FPO, CPO, HPO, SEO, and AEO contained 37, 39, 41, 27, and 39 VOCs, respectively. This demonstrated that extraction methods differ in their effects on retaining and releasing VOCs in IPOs (Zhang et al., 2025a). Nine common flavor components were identified across all IPOs and were key contributors to their characteristic flavor profile. Among the 9 VOCs, α–lonone exhibits a violet aroma, is abundant in plant essential oils, and is widely used in flavor and fragrance applications. Phenethyl alcohol, an edible flavoring agent naturally present in essential oils, possesses a rose–like scent. Together with linalool, eugenol, and other aromatic compounds, it constitutes the natural aroma profile of rose essential oil. However, eugenol suffers from poor thermal stability and high solubility in organic solvents. Consequently, it had only been detected in FPO and SEO. Additionally, methyl salicylate and guaiacol are two volatile phenolic compounds widely found in woody plant essential oils, imparting minty and woody aroma to IPOs, respectively. 1,2–cyclohexanedione contributed sweet caramel and nutty peel flavors to IPOs, further enriching their overall aroma characteristics.

Fig. 2.

Heat map of volatile components of five IPOs. The redder color represents the relatively higher compound content, and the bluer color represents the relatively lower compound content

The content of VOCs in the IPOs was represented as a stacked area chart by category as shown in Fig. 3a, and significant differences were observed in the proportions of various types of compounds. It is worth noting that phenolic and aldehyde VOCs were notably prominent in the five IPOs. This prominence was linked to their high volatility and abundance in IPOs, directly reflecting the characteristic flavor profile. Ketones (38.25%) and phenols (28.40%) were the major components identified in the FPO. In contrast, volatile alcohols constituted the largest proportion (36.79%) in the CPO. For the HPO, the combined content of volatile alcohols (29.78%) and aldehydes (23.99%) accounted for approximately 53.77% of the VOCs. Hexanal and benzaldehyde, the main representatives of these VOCs, constituted and coordinated aroma profile of HPO, imparting grassy, apple, and almond notes (Grebenteuch et al., 2021; Majcher et al., 2025). Ester compounds, primarily produced via esterification reactions, constitute the highest proportion (26.20%) in SEO (Zhang et al., 2020). This suggested that the relatively mild solvent extraction conditions favor the formation and retention of these esters. Aldehyde compounds, including valeraldehyde, salicylaldehyde, and benzaldehyde, accounted for 22.73% of AEO. It is worth noting that valeraldehyde has a fermented bread and fruity aroma and is used in the production of edible spices such as flavor enhancers (Zhang et al., 2022).

Fig. 3.

Stacked plots of compound contents (a), OPLS–DA (b), displacement test result (c), VIP value (d) of volatile components of five IPOs. In the stacked diagram, different colors represent different types of volatile compounds

The Orthogonal Partial Least Squares–Discriminant Analysis (OPLS–DA) model enabled the discrimination of flavor similarities and differences among the IPOs. In this study, the model demonstrated excellent goodness–of–fit and statistical significance. The independent variable fit index (R²X), the dependent variable fit index (R²Y), and the predictive capability (Q²) were 0.971, 0.991, and 0.981, respectively. As shown in the score scatter plot (Fig. 3b), distinct clustering trends and clear separation among IPOs were observed, with the cumulative variance contribution rates of PC1 and PC2 reaching 77.5%, indicating high reliability of the overall dataset and robustness of the analytical model. It can be observed that the five IPOs were well separated on PCA1 and PCA2, indicating that PCA effectively captured the main differences between IPOs (Miao et al., 2024). The closer proximity of CPO and HPO, as well as FPO and AEO, in the plot suggested their similar aromatic profiles. Conversely, SEO exhibited significant divergence in flavor composition compared to the other IPOs, which is consistent with the hierarchical clustering heatmap (Fig. 2). A permutation test consisting of 200 iterations (Fig. 3c) verified the validity of the model, and the Q²–regression line intersected the y–axis below zero, ruling out overfitting. By calculating the Variable Importance in Projection (VIP) value, the influence and discriminatory power of each VOC on the classification differences between IPOs can be assessed. A higher VIP value indicates a greater contribution of that VOCs to the aroma profile differences among IPOs, signifying its importance in distinguishing between extraction methods. Based on the VIP > 1.0 threshold, 14 VOCs were identified as key differential marker compounds for the five IPOs (Fig. 3d). They were mainly alcohols, phenols, aldehydes and ketones. Among them, linalool is a terpene alcohol that imparted rich citrus, floral, and rose–like sweetness to IPOs (Wang et al., 2025). In summary, different extraction methods had a significant impact on the composition of VOCs in IPOs and made unique contributions to their flavor characteristics.

Infrared spectral

The infrared spectral scanning results of IPOs are presented in Fig. S1a, which revealed highly similar waveforms across the five IPOs and disclosed intermolecular information associated with the fundamental vibrations of functional groups. Infrared spectral band assignments of IPOs are documented in the Supplementary Materials.

Rheological property

As depicted in Fig. S1b, the shear rate–shear stress curves of the IPOs exhibited an excellent linear relationship. This behavior, closely linked to the composition and structure of long–chain molecules and fatty acids in vegetable oils, indicates that the IPOs conform to Newtonian fluid behavior (Zhang et al., 2024). The viscosity of each IPO was determined by fitting the shear stress–shear rate curve (Table S2), revealing significant disparities (52.53–60.53 mPa·s). Specifically, SEO exhibited the lowest viscosity (52.53 mPa·s), while FPO had the highest (60.53 mPa·s). The relatively higher extraction temperature used in hot pressing may have generated polymers and dimers, explaining why HPO viscosity (57.90 mPa·s) exceeded that of cold–pressed CPO (55.97 mPa·s). Furthermore, β–sitosterol can form microcrystalline or micellar structures at low temperatures, impeding the free flow of lipid molecules and thereby increasing viscosity (Suri et al., 2022). The observed association between sterol content and viscosity in IPOs substantiates this mechanism. Previous study has shown that the viscosity correlates with oleic acid and linoleic acid content (Musaözcan et al., 2019). The relationship established between viscosity and oleic acid content in FPO is consistent with this established rule.

Antioxidant activity

DPPH radical scavenging capacity

As shown in Fig. S2a, d, the DPPH radical scavenging capacity of the IPOs and Vc increased linearly with concentration. The corresponding IC₅₀ values (half–maximal inhibitory concentration) are presented in Table S3. Based on these IC₅₀ values (ranging from 2.82 to 8.01 mg/mL), the IPOs demonstrated significant scavenging activity. Notably, the IC₅₀ for SEO was significantly lower than that repor0ted by Kagliwal et al. (2012) for seabuckthorn berry oil extracted via supercritical CO₂ (29.02 mg/mL). Among the IPOs, AEO exhibited the strongest DPPH radical scavenging capacity, which can be attributed to its elevated concentration of phenolic compounds. The hydroxyl groups in these phenolics exhibit significantly enhanced scavenging activity against DPPH radical (Likhar et al., 2025). At a concentration of 12 mg/mL, AEO achieved a DPPH radical scavenging rate of 99.71%.

ABTS radical scavenging capacity

Fig. S2b, e shows the scavenging efficacy of IPOs and Vc against ABTS radical. Their ABTS radical scavenging capacity exhibited a clear concentration–dependent trend. The IC₅₀ values of the five IPOs (FPO, CPO, HPO, SEO, AEO) were 31.33, 17.37, 20.72, 20.62, and 17.30 mg/mL, respectively. AEO had the lowest IC₅₀ value for ABTS radical scavenging, further confirming its superior antioxidant capacity. This may be attributed to the high content of hydrophilic compounds in AEO, which efficiently inhibited the progression of the oxidation reaction (Ponnusamy et al., 2024). Additionally, the IC₅₀ values of each IPO for scavenging DPPH radical were significantly higher than those for scavenging ABTS radical (Table S3), indicating that the IPOs were less effective at scavenging ABTS radical compared to DPPH radical.

Total reducing power

The reducing potential of antioxidants is typically determined by their ability to convert ferric ions (Fe³⁺) to ferrous ions (Fe^2*) in acidic media (Shahidi and Zhong, 2015). As shown in Fig. S2c, f, the reducing power increased rapidly with increasing concentrations of IPO and Vc. This reducing power was further quantified by comparing RP_0.5AU values (the IPO concentration required to reach an absorbance of 0.5 at 700 nm). As indicated in Table S3, the RP_0.5AU value was the lowest for AEO (15.30 mg/mL), followed by HPO (15.97 mg/mL) and CPO (21.65 mg/mL). In contrast, Vc achieved the same absorbance at a significantly lower concentration (0.20 mg/mL). These results demonstrates that among the IPOs tested, AEO exhibited the strongest reducing capacity. As reported by Buamard et al. (2023), the remarkable reducing ability of AEO may be partly attributed to its high content of polyphenols and flavonoids.

In conclusion, this study demonstrated that FPO exhibits the lowest acid and peroxide values while retaining higher concentrations of tocopherols, squalene, β–sitosterol, and carotenoids. The high linoleic acid content in IPO provided new opportunities for functional product development. Crucially, this work provided the following operational guidance for the factories. Fresh pressing extraction maximizes retention of heat–sensitive micronutrients within oil crops. Both cold pressing and hot pressing extraction methods can promote the dissolution of flavor compounds, exerting different effects on the quality of the oil. Solvent extraction may be promising method for large–scale oil extraction from the perspective of yields. Aqueous extraction offers an eco–friendly approach to concentrate polyphenols and flavonoids as natural antioxidants. Future research should focus more on emerging oil extraction technologies and clarify the connection between extraction methods and oil quality by establishing databases. Additionally, it should optimize processing through comprehensive analysis of the flavor compounds in the oil to ensure its nutritional quality and safety.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

This work is supported by the National Key Research and Development Program of China (Grant No. 2023YFE0116100); National Natural Science Foundation of China (Grant No. 32472256).

Declarations

Conflict of interests

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