Headspace stir‐bar sorptive extraction combined with gas chromatography–mass spectrometry for trace analysis of volatile organic compounds in Schisandra chinensis Baillon (omija)

Abstract Analyzing volatile organic compounds (VOCs) in food is crucial but challenging. Schisandra chinensis Baillon (omija) is an herbal plant with various functional health activities. Previous VOC analyses focused on S. chinensis fruit but not its leaves. Therefore, VOCs in S. chinensis fruit and leaves were analyzed using headspace stir‐bar sorptive extraction (HS‐SBSE)‐GC–MS, and optimal conditions were established. Various factors, such as the sample preparation method, twister stir‐bar type, sample amount, extraction temperature, and extraction time, expected to affect extraction were carefully optimized. Under the optimal conditions, 35 and 40 VOCs were identified in S. chinensis fruit and leaves, respectively. This HS‐SBSE method is capable of rapid analysis and a low contamination rate without requiring organic solvents. These findings provide practical guidelines for HS‐SBSE applications in various food matrices by providing analytical methods for VOC detection.

| 7397 have a significant impact on the odor activity of food (Lee et al., 2019).
However, it is challenging to accurately analyze VOC profiles in food because of their abundance and the fact that the concentrations of each VOC vary (Lee et al., 2019).Analysis of VOCs from samples is carried out in several steps, including sampling, sample extraction, separation, detection, and analysis (Marín-San Román et al., 2020).Sample extraction is a very important step that can lead to serious errors owing to the loss of a large amount of analyte (Andrade-Eiroa et al., 2016).
Therefore, many sample extraction methods have been developed and are constantly being modified, and new methods have been developed to achieve optimal analytical results (Lee, Cha, et al., 2021;Marín-San Román et al., 2020;Serrano de la Hoz et al., 2016).Liquid-liquid extraction, solid-phase extraction, simultaneous distillation extraction, and distillation under reduced pressure, which are well-known conventional sample extraction techniques, have been reported to have several problems, such as low reproducibility, low selectivity, and difficult automatization (Marín-San Román et al., 2020;Song et al., 2021).
Moreover, these conventional extraction methods use a large amount of solvent, which increases the risk of workplace and environmental pollution (Silvestre et al., 2009).Solid-phase microextraction (SPME) and stir-bar sorptive extraction (SBSE) have been developed as new extraction methods to solve these problems.Both methods eliminate the use of solvents and are relatively simple-extraction and concentration are performed in one step (Ochiai et al., 2018;Perestrelo et al., 2011).SBSE has a smaller sample loss volume than SPME and is more sensitive and robust than SPME (Lee et al., 2018;Prieto et al., 2010).According to previous reports, SBSE is 50-250 times more sensitive than SPME (Marín-San Román et al., 2020).The SBSE technique uses a magnetic bar, called a "twister," to extract the sample.This magnetic bar, which is coated with a polymeric extracting phase, extracts and enriches organic compounds from aqueous matrices (David & Sandra, 2007).Headspace-SBSE (HS-SBSE) is an analytical method that can be applied to all solid, liquid, and gaseous samples, in which the twister is introduced into a vial adapted for headspace (Marín-San Román et al., 2020).Headspace sampling has the advantage of very high selectivity because only volatile and semi-volatile organic compounds are released into the headspace.In addition, because there is no contact between the sample and the phase, both the background adsorption and matrix effect are reduced, and the life expectancy of the bar is increased (Grossi et al., 2008).To the best of our knowledge, only a few studies have identified VOCs in the fruit and leaves of S. chinensis using the HS-SBSE extraction method.Therefore, this study aimed to develop and validate an optimized analytical method for the characterization and determination of VOCs in the fruit and leaves of S. chinensis using HS-SBSE and GC-MS.

| Materials and reagents
To determine the sample pretreatment method to extract volatile compounds from the fruit and leaves of S. chinensis, raw fruit and leaf samples were freeze-dried at −70°C for 53 h, and then the ratio of the raw samples to the retrieved ones was measured.
The retrieved samples of equivalent weight to the raw samples were used as extraction samples.The amounts of retrieved fruit and leaf samples after freeze drying were 18.07% ± 0.17% and 17.16% ± 0.94% of the raw samples, respectively.The raw samples, blended raw samples, and freeze-dried blended samples were placed in a Twister headspace vial (Gerstel) and used for the comparative analysis of the sample pretreatment methods to extract VOCs.The fruit and leaves of S. chinensis were obtained from Hyojongwon, a local producer located in Munkyung, Gyeosangbukdo, Korea.All S. chinensis samples were stored in a freezer at −20°C until analysis.Before HS-SBSE, all samples were ground with dry ice to prevent the thermal loss of VOCs.All chemical reagents used in this study were purchased from Sigma-Aldrich Corporation (St. Louis, MO, USA).An internal standard solution (phenethyl alcohol) was prepared by dilution in distilled water.

| Optimization of HS-SBSE
To establish the optimal conditions for analyzing VOCs in the fruit and leaves of S. chinensis using HS-SBSE, different preparation methods, twister stir-bar types, sample amounts, extraction temperatures, and extraction times were evaluated.The preparation methods for the analysis included raw samples, blended raw samples, and freeze-dried blended samples.We tested two different stir bars, comprising either polydimethylsiloxane (PDMS) or ethylene glycol (EG) silicone.The tested sample amounts were 0.3, 0.6, and 0.9 g for fruit and 0.1, 0.2, 0.4, and 0.6 g for leaves.The tested temperatures were 30, 40, and 50°C, and the tested extraction times were 30, 60, 90, 120, and 150 min.Finally, analysis was performed using GC-MS under the conditions mentioned in the next section, and the optimal conditions were selected by comparing the adsorption efficiencies.

| GC-MS analysis
For the extraction of S. chinensis, fruit (0.3 g) and leaves (0.4 g) were placed in Twister headspace vials (Gerstel) before 100 μL of phenethyl alcohol (10 g/L) solution was added as an internal standard.
Phenethyl alcohol was chosen as the internal standard after confirming its absence in S. chinensis samples and its effective separation from other VOCs under optimized conditions.A Twister headspace insert (Gerstel) was placed in the upper part of the vial.A twister was then placed in the insert, and the vial was sealed with a crimp cap (Gerstel).For extraction, a vial containing a stir bar was placed in an agitator or a heating block.After extraction, the stir bar was placed in a Twister desorption liner (Gerstel) before the analytes on the stir bar were thermally desorbed in a thermal desorption unit.VOCs were identified in S. chinensis samples, together with their relative peak areas and RIs on DB-WAX.The quantification of VOCs was performed using the peak area ratio (peak area of each compound/peak area of the internal standard).Due to the absence of standard curves for individual characteristic volatile compounds, semi-quantitative determinations were carried out with phenethyl alcohol as the internal standard.While no standard curves were available, the use of relative concentrations remains valuable for the analysis of volatile profiles in foods (Jang et al., 2021;Lee et al., 2021;Xiao et al., 2014;Yao et al., 2015).Therefore, in this study, the relative contents of VOCs are expressed as equivalents of the internal standard (100 μL of 10 g/L phenethyl alcohol) in the HS-SBSE-GC-MS analysis.All experiments were performed in triplicate.The tentative identification of VOCs was performed by comparing the retention indices and mass spectra with those of the mass spectral library.The experimental retention index of VOCs in S. chinensis samples was calculated using saturated alkane standards containing C7-C40 (Sigma-Aldrich, Steinheim, Germany), and the mass spectra of each VOC were compared to the information obtained from the Wiley and NIST 08 library databases.The extracted VOCs were analyzed using an HP 6980 gas chromatograph coupled with an HP 5973 mass-selective detector (Agilent Technologies).The gas chromatograph was equipped with a 60-m DB-WAX column (internal diameter = 0.25 mm and thickness = 0.25 μm; Agilent Technologies).The temperature of the thermal desorption unit was programmed to increase from 50 to 220°C at 60°C/min for 5 min.The desorption flow in the thermal desorption unit was maintained at 50 mL/min in splitless mode.The temperature of the cooled injection system was maintained at −20°C with liquid nitrogen gas, and it was then programmed to 220°C at 12°C/s (held for 2 min) in 20:1 split mode, with a helium gas flow rate of 1.4 mL/min.The oven of the GC was maintained at 50°C (1 min) and then programmed to reach 210°C at 3°C/min.The transfer line, ion source, and quadrupole were maintained at 250, 230, and 150°C, respectively.The mass spectra were employed in full-scan mode, and the mass range was collected between 35 and 400 m/z.

| Statistical analysis
All experiments were performed in triplicate, and the data are presented as the mean ± SD.Data were statistically analyzed using oneway analysis of variance (ANOVA) with Duncan's multiple-range test (p < .05)and student's t-test in SPSS Statistics 20 (SPSS Inc.).

F I G U R E 1
Effect of the sample preparation method using headspace stir-bar sorptive extract-GC-MS for the extraction of volatile organic compounds in Schisandra chinensis fruit (A) and leaves (B).Values are expressed as the mean ± SD.Different letters (a-c) among samples indicate significant differences calculated via one-way ANOVA followed by Duncan's multiple-range test (p < .05).Normalized peak area (%) = peak area of sample/the highest peak area of sample ×100.

| Optimization of HS-SBSE extraction conditions
To establish the optimal conditions for the analysis of VOCs in S. chinensis fruit and leaves, the factors affecting extraction efficiency, including the sample preparation method, type of stir bar, sample amount, extraction temperature, and time, were established.

| Effect of sample preparation
To confirm the optimal conditions for the sample preparation method, three types of samples were prepared and analyzed: raw samples, blended raw samples, and freeze-dried blended samples.As shown in Figure 1, eight VOCs (α-ylangene, p-cymene, γ-terpinene, α-terpinene, β-myrcene, sabinene, β-pinene, and βhimachalene) were commonly detected as the major components of S. chinensis fruit samples prepared by the three methods.Similarly, it was confirmed that six major VOCs (γ-terpinene, germacrene D, (-)β-elemene, (E)β-ocimene, sabinene, and β-pinene) were commonly detected in the leaf samples of S. chinensis prepared by the three methods.When the extraction efficiencies of the three preparation methods were compared based on the peak area of major VOCs, it was confirmed that the freeze-dried blended sample had a high extraction efficiency for both fruit and leaves (p < .05).Thus, the freeze-dried blended method was better suited for extraction efficiency.that acetic acid and α-thujene were detected only with EG-silicone.In contrast, in the leaf samples, major VOCs, including (E,E)α-farnesene, (-)β-elemene, (Z)-3-hexen-1-ol, 1-hexanol, and 1-penten-3-ol, were detected with both stir bars.However, when the EG-silicone stir bar was used, the extraction efficiency was higher based on the peak area (p < .05 or .01).This result is considered to be because EG has a higher affinity for polar substances than PDMS (Marín-San Román et al., 2020).Thus, the use of an EG-silicone stir bar for HS-SBSE extraction of VOCs from S. chinensis fruit and leaves is recommended.

| Effect of sample amount
To optimize the sample weight of freeze-dried blended S. chinensis fruit and leaves using HS-SBSE, VOCs were extracted at 50°C for 2.5 h with an EG-silicone stir bar.As shown in Figure 3, the extraction efficiencies of methyl carvacrol, p-cymene, sabinene, β-pinene, and β-himachalene, which are the main VOCs of S. chinensis fruit, were compared by weight, and there was no significant difference (p > .05) in the extraction efficiency according to the weight of the sample.Therefore, 0.3 g was considered the most suitable for extracting S. chinensis fruit samples.In S. chinensis leaves, it was confirmed that the extraction efficiency increased significantly as the amount of sample increased from 0.1 to 0.4 g (p < .05)for the six main VOCs (γ-terpinene, germacrene D, (-)β-elemene, (E)β-ocimene, sabinene, and β-pinene).However, there was no significant difference between 0.4 and 0.6 g of the sample (p > .05);thus, the optimum sample weight for analysis of VOCs from S. chinensis leaves was estimated to be 0.4 g.

| Effect of extraction temperature
The optimal extraction temperature for obtaining the highest extraction efficiency of VOCs from S. chinensis fruit and leaves using

| Effect of extraction time
Extraction time is an important parameter that affects the extraction process of headspace sampling (Zhao et al., 2009).The optimal extraction time for obtaining the highest extraction efficiency was confirmed by performing extraction for 30, 60, 90, 120, and 150 min.

| Selection of the internal standard
To evaluate the internal standard for quantifying the VOCs of S. chinensis using HS-SBSE, eight internal standards, including phenethyl alcohol, methyl octanoate, tetradecane, ethyl nonanoate, 1-hexanol, hexyl butanoate, hexyl 2-methyl butanoate, hexyl isobutyrate, and hexyl hexanoate, were tested.As a result, the peaks of seven internal standards, except for phenethyl alcohol, tended to overlap with those of VOCs extracted from S. chinensis fruit and leaves or affected the total peak area.In addition, an internal standard with F I G U R E 4 Effect of extraction temperature using headspace stirbar sorptive extract-GC-MS for the extraction of volatile organic compounds in Schisandra chinensis fruit (A) and leaves (B).Values are expressed as the mean ± SD.Different letters (a-c) among samples indicate the significant differences calculated via one-way analysis of variance followed by Duncan's multiple-range test (p < .05).Normalized peak area (%) = peak area of sample/the highest peak area of sample ×100.
high water solubility was prioritized to take advantage of the environmentally friendly HS-SBSE, which does not use a solvent.Therefore, phenethyl alcohol was selected as the internal standard for the qualitative and quantitative analyses in this study.

| Analysis of VOCs in S. chinensis fruit and leaves using HS-SBSE under optimal conditions
The optimal conditions for analyzing VOCs in S. chinensis fruit and leaves by HS-SBSE were as follows: sample preparation method = freeze-dried blended; twister stir-bar type = EG-silicone stir bar; sample amount = 0.3 g (fruit) and 0.4 g (leaves); extraction temperature = 50°C; and extraction time = 150 min (fruit) and 120 min (leaves).Under the optimized extraction conditions, the volatile compounds in S. chinensis samples were evaluated using HS-SBSE.In total, 56 VOCs were identified in all S. chinensis samples (Table 1).
A total of 35 VOCs were identified from the extracts of S. chinensis fruit: 15 monoterpene hydrocarbons, 8 sesquiterpene hydrocarbons, 2 oxygenated terpenes, 4 ketones, 2 hydrocarbons, 1 ether, 1 ester, 1 aldehyde, and 1 acid compound.Quantitatively, the main volatile components of S. chinensis fruit were monoterpenes and sesquiterpenes, including β-myrcene, α-terpinene, limonene, γ-terpinene, p-cymene, α-ylangene, and β-himachalene; these findings are similar to those of a previous study (Lee et al., 2011).Kim et al. (2008) also reported that β-myrcene, α-terpinene, limonene, γ-terpinene, and p-cymene are the major compounds in the fruit of S. chinensis.γ-Terpinene, a major compound in fruit, is known to effectively inhibit lipid oxidation, which is the main component of cell membranes (Guo et al., 2021).In addition, the biofunctional activities of β-myrcene, which is present in the fruit and leaves, such as antioxidant and antibacterial activity, have been reported (Wang et al., 2019).According to a previous study (Wang et al., 2019), β-myrcene, along with limonene, another major compound in fruit, has antibacterial activity against various foodborne pathogenic F I G U R E 5 Effect of extraction time using headspace stir-bar sorptive extract-GC-MS for the extraction of volatile organic compounds in Schisandra chinensis fruit (A) and leaves (B).Values are expressed as the mean ± SD.Different letters (a-d) among samples indicate the significant differences calculated via one-way analysis of variance followed by Duncan's multiple-range test (p < .05).Normalized peak area (%) = peak area of sample/the highest peak area of sample ×100.

TA B L E 1
Volatile organic compounds of fruit and leaves of Schisandra chinensis obtained using headspace stir-bar sorptive extract-GC-MS.(Teng & Lee, 2014).These were the major compounds found in the fruit of S. chinensis.
The following 40 VOCs were identified from the extracts of S. chinensis leaves: 15 monoterpene hydrocarbons, 14 sesquiterpene hydrocarbons, 1 oxygenated terpene, 1 ketone, 1 ester, 5 aldehydes, and 3 alcohol compounds.Quantitatively, the main volatile components in S. chinensis leaves included sabinene, β-myrcene, (E)β-ocimene, (-)β-elemene, and germacrene D, similar to those found in a previous study (Zheng et al., 2005).Among the major compounds of S. chinensis leaves, germacrene D is also found in the leaves of various other plants and is characterized by its excellent antioxidant activity (Andrade-Eiroa et al., 2016;Xie et al., 2015).In addition, (E)β-ocimene is found in the leaves and stems of Gypsophila bicolor, and it has been reported to have antibacterial activity against Gram-positive and Gram-negative bacteria (Shafaghat & Shafaghatlonbar, 2011).In summary, the major compounds detected in the fruit and leaves of S. chinensis showed excellent biofunctional activity as well as aromatic components, suggesting that they are highly likely to be used in the food industry.
A total of 35 VOCs were detected in S. chinensis fruit, and 40 VOCs were detected in leaves, confirming that more VOCs were detected in the leaves.In addition, 19 VOCs were simultaneously detected in both the fruit and leaves.While the total concentration of VOCs was higher in the fruit, the leaves had 21 types of VOCs that were not found in the fruit.The leaf-specific VOCs included (E)β-ocimene, δ-elemene, and (-)β-elemene, which were characterized by a pleasant odor.
Until now, the SBSE method has mainly been used to analyze liquids, such as beer, wine, juice, and milk.The use of the SBSE method for the analysis of solid foods, such as fruit and vegetables, is only possible after extraction with an organic solvent.However, in this study, the VOCs of solids (fruit and leaves) were successfully analyzed without using organic solvents through the HS-SBSE method.
The optimal conditions for analyzing VOCs in S. chinensis fruit and leaves using HS-SBSE were confirmed through our experiments, and under these conditions, 35 and 40 VOCs were identified from S. chinensis fruit and leaf extracts, respectively.Thus, an analysis method with excellent reproducibility that reduces the number of by-products and the rate of contamination of the sample was established in this study; furthermore, the method is fast and does not require organic solvents.These results support the use of HS-SBSE for the analysis of VOCs in solid foods and are expected to improve food research in the future.TA B L E 1 (Continued) 3.1.2| Effect of the twister stir-bar type After choosing the freeze-dried blended S. chinensis fruit and leaves in a headspace vial, extraction was performed using PDMS and EGsilicone stir bars.In the fruit samples (Figure 2), major VOCs, such as α-ylangene, p-cymene, γ-terpinene, sabinene, β-pinene, and βhimachalene, were detected with both the PDMS and EG-silicone stir bars.When the EG-silicone stir bar was used, the extraction efficiency of several VOCs was lower than that of PDMS one, but it was confirmed F I G U R E 2 Effect of the twister stir bar using headspace stir-bar sorptive extract-GC-MS for the extraction of volatile organic compounds in Schisandra chinensis fruit (a) and leaves (b).Values are expressed as the mean ± SD.Statistical differences are indicated by *p < .05,**p < .01,and ***p < .001(Student's t-test) for comparisons between EG-silicone and polydimethylsiloxane stir bars; ns: not significant.Normalized peak area (%) = peak area of sample/the highest peak area of sample ×100.
HS-SBSE was examined by performing extraction at 30, 40, and 50°C.As shown in Figure 4, the peak areas of the VOCs (α-ylangene, p-cymene, γ-terpinene, α-terpinene, β-myrcene, sabinene, β-pinene, and β-himachalene) at 50°C were the highest among different F I G U R E 3 Effect of sample amount using headspace stir-bar sorptive extract-GC-MS for the extraction of volatile organic compounds in Schisandra chinensis fruit (A) and leaves (B).Values are expressed as the mean ± SD.Different letters (a-c) among samples indicate the significant differences calculated via one-way analysis of variance followed by Duncan's multiple-range test (p < .05).Normalized peak area (%) = peak area of sample/the highest peak area of sample ×100.temperatures in the case of S. chinensis fruit (p < .05).Similarly, the peak areas of the VOCs ((E)-2-hexenal, (Z)-3-hexen-1-ol, germacrene D, (E)β-ocimene, sabinene, and β-pinene) from S. chinensis leaves considerably increased until extraction reached 50°C.Therefore, to achieve a high extraction efficiency of S. chinensis fruit and leaves, the extraction temperature was set to 50°C.

Note:
Abbreviations: -, not detected; RI, retention index; RT, retention time.a Identification of VOCs from the acquired spectra was achieved based on a high match factor (>800).b Relative concentration (mg/kg) was obtained by the following formula: Sample concentration = Internal standard concentration (mg/kg) × Sample peak area/Internal standard peak area.c Odor description from: https://www.vcf-online.nl/VcfHome.cfm.