Essential Oil Extracted from Cymbopogon citronella Leaves by Supercritical Carbon Dioxide: Antioxidant and Antimicrobial Activities

To improve essential oil quality, especially to reserve the thermal instability of compounds, supercritical CO2 extraction (SFE) was applied to recover essential oil from Cymbopogon citronella leaves. A response surface methodology was applied to optimize the extraction process. The highest essential oil yield was predicted at extraction time 120  min, extraction pressure 25  MPa, extraction temperature 35°C, and CO2 flow 18  L/h for the SFE processing. Under these experimental conditions, the mean essential oil yield is 4.40%. In addition, the chemical compositions of SFE were compared with those obtained by hydrodistillation extraction (HD). There were 41 compounds obtained of SFE, while 35 compounds of HD. Alcohols and aldehydes were the main compositions in the essential oils. Furthermore, the antioxidant activities and antimicrobial of essential oils obtained by HD and the evaluated condition of SFE were compared. Results showed that the antioxidant activities of SFE oil are better than those of HD. Minimum inhibitory concentrations (MICs) were determined by the microdilution method. Essential oil obtained from SFE and HD exhibited a significant antimicrobial activity against all tested microorganisms. It is confirmed that the SFE method can be an alternative processing method to extract essential oils from Cymbopogon citronella leaves.


Introduction
Cymbopogon citronella Stapf (C. citronella), also known as lemongrass, belongs to Cymbopogon Spreng. family. C. citronella is a perennial grass that is widely cultivated in tropic and subtropic regions of the eastern hemisphere, such as China [1,2]. C. citronella is barren and drought resistant and can tolerate a temperature range from −25°C to 40°C, suggesting that it can be planted in a wide range of environmental conditions [3]. e essential oil of C. citronella has been studied widely; Sriramavaratharajan et al. [3] has found that the essential oil was mainly distributed in the C. citronella leaves. In other research studies, the components of the C. citronella essential oil were identified, including mainly citral (aldehydes geranial + neral) and terpenes (myrcene-monoterpene and geranial-terpenic alcohol) [1,2]. As an economic oil crop, the fragrance odour together with the various pharmacological activities promote the essential oil widely used as a flavor enhancer in food, perfumery, soap, cosmetic, pharmaceutical, insecticide industries, etc. [4,5]. For example, the citral from the essential oil could be used as a raw material to prepare vitamins (A, E, and K), ionone, methylionone, and perfumes [1,6]. Consequently, this essential oil-producing crop has attracted huge economic interests in China and abroad in recent years.
However, the C. citronella essential oil has always been obtained by hydrodistillation extraction (HD) method. Actually, the HD is a widely used method to measure the yield of thermally stable essential oil even today. Various researchers have extracted the essential oil from different oil crops with HD [1,2]. Other studies [1,2] extracted the essential oil (of 0.03% feedstock) from Artemisia capillaris T. by HD. Huang [7] extracted 0.25% and 1.75% of the essential oil from Cinnamomum cassia P. and Zingiber officinale R. by the HD. 2.69% of essential oil was extracted from C. citronella by the HD with the response surface method (RSM) reported in [1,2]. In addition, Sargenti and Lanças [2] performed and reviewed the relative analysis for extraction efficiently of essential oil with various methods, indicating HD can be widely used as a common processing for most thermally stable essential oils. However, many research studies had reported that the high temperature and long treatment time in extraction processing contributed to degradation of thermosensitive compounds and need a great amount of energy consumption [8]. us, it is necessary to continue to search and establish possible alternative process to further improve essential oil quality. Furthermore, in considering the subsequent separation of the essential oil in extraction processing, our group proposed the method of supercritical fluid extraction (SFE) in this study.
As a newly and environmentally separating technology, SFE, also referred to as pressurized extraction, employs environmental-friendly agents, such as CO 2 and water, as the mediator. SFE is an effective and selective sample separation technique, which has been widely used in industrial production, environmental analysis, and analytical chemistry [9,10]. Several studies have demonstrated that the SFE process was applied in extracting bioactive compounds, such as flavonoids, active polysaccharide, protein, polyphenol, and essential oil [9]. Compared with various other extraction methods, Pavela [1] and SargentiLanças [2] found that the SFE was efficient in selectively fractionating the essential oil under mild extraction conditions. ey also evaluated the effect of different factors on extraction yield of essential oil by SFE, resulting in 1.51% of essential oil [1,2]. To our knowledge, nevertheless, there was no report in the literature on using SFE in extraction of essential oil from C. citronella.
In addition, antioxidant and antimicrobial chemistry are interesting topics at present because they play a pivotal role in pathogenesis of cardiovascular diseases, neural disorders, cancer, and aging [11]. Consequently, a majority of the reports also have focused on biological characteristics and compositions of essential oil extracted from various other plants [4,8,11,12], such as Rosmarinus officinalis, A. capillaris, and Z. officinale, but there was few information about the essential oil of C. citronella leaves essential oil in the literature. e objective of this study was therefore to optimize the SFE conditions so as to obtain the highest essential oil yield from C. citronella leaves. In this study, the supercritical carbon dioxide was employed to extract the essential oil, and the operational parameters were optimized using response surface methodology. e chemical compositions of the essential oil obtained from C. citronella leaves were also identified by gas chromatography mass spectrometry (GC-MS) analysis. Furthermore, the antioxidant activities and antimicrobial activities of the essential oil extracted by SFE were compared with those obtained by HD.

Materials and Methods
2.1. Materials. Fresh C. citronella leaves were collected from a local farm in Changsha, Hunan, soon after harvesting, and then transported and preserved in our laboratory. e leaves were air-dried (6% of moisture) and milled into powders using a frozen broken molecular machine (SPEX Sample-Prep, USA). Powders were screened through an 80-mesh sieve and kept in a refrigerator at 4°C until further use.

Hydrodistillation (HD)
Process. 100 g of dried milled powder was immersed in 300 mL water and distilled for 3 h, using a Clevenger-type apparatus which was found to be sufficient for completing the extraction process. e essential oil was collected and was then weighed. e collected oil was dried with a small amount of anhydrous sodium sulfate and refrigerated prior to further analysis.

Supercritical Fluid Extraction (SFE) Process.
e extraction processing was conducted in a 500 mL extraction vessel (Hunan Supercritical Extraction Company Ltd., Jiangsu province). e extraction was according to S. R. Sargenti [2] with minor modification. Briefly, a 200 g dried substance would be placed in an extraction kettle, including extractor and separator would be heated to 30°C and 25°C, respectively. CO 2 is cooled in a condenser and filled into the extractor. e pressurization was conducted to maintain the pressure at 20 MPa by filling the CO 2 ; thereafter, the flow of CO 2 was adjusted to 20 L/h for essential oil by circulatory extraction. After extraction, the separation temperature and pressure were controlled to 25°C and 4 MPa, respectively. To evaluate the effect of the variables on the extraction yield of essential oil, those factors, including the extraction time, extraction temperature, and CO 2 flow, were studied, respectively, with single-factor experimental. e standard deviation is less than 4.0%.

Optimization of SFE.
Based on the single-factor study, the optimized experimental conditions for maximum extraction of essential oil from C. citronella leaves using supercritical CO 2 were determined using the design-expert V8.0.6. Given the influence of the above variables, the extraction processing of essential oil was optimized with a Box-Behnken design (BBD) and response surface methodology (RSM) [13]. For the BBD experiments of 4 factors, there were 24 experiments augmented and 3 replications at the center values (zero level). e levels of each factor and the design table are given in Tables 1 and 2. Design-expert V8.0.6 was used for the experiment designs and subsequent regression analysis of the response data. Statistical analysis of the model was performed to evaluate the analysis of variance (ANOVA). e quality of the polynomial model equation was judged statistically using the coefficient of determination R 2 and adjustment R 2 , and its statistical significance was determined by the F value and P value. e significance of the regression coefficients was tested by some parameters, such as coefficient of variation (CV) and adequate precision.

GC-MS Analysis.
e compositions of essential oil obtained by HD and SFE (in evaluated condition) were analyzed by a GC-MS analysis. e examination was performed on a Scion SQ (Bruker, USA) equipped with an 8400 autosampler. Samples were injected into the capillary column in the split and splitless modes. A 1 μL aliquot of test solution was injected into a DB-5 (30 m × 250 μm × 0.25 μm, Agilent, USA). e oven temperature was programmed at 5°C/min from 40°C (hold 2 min) to 100°C (hold 2 min) and then 10°C/min to 250°C (hold 2 min). e injector temperature was set at 230°C in the split and splitless modes. e flow rate of the carrier gas helium was set at 1 mL/min, the transmission line temperature was set at 200°C, and the ion source temperature was set at 250°C. e electron ionization (EI) source operated at −70 eV. e MS spectra were monitored in the full scan mode from 15 to 500 m/z with a scan event time of 0.3 s and a scan speed of 2000 μ/s. e solvent delay time was set to 6.00 min. MS data analyzed by GC/MS were processed by Analyzer Pro Software (Spectralworks Ltd., Runcorn, UK) including deconvolution of mass spectra, data collection, alignment, and normalization. Metabolite peaks were deconvoluted and collected using an area threshold of 7,000, height threshold of 1, signal to noise ratio of 3, width threshold of 0.01, scan windows of 5, and smoothing of 5. e data were aligned with a retention time window of 0.3 min.

Self-Oxidation Assay.
e scavenging ability of 1,2,3phentriol to self-oxidize was investigated according to the method with a minor modification [14]. AA and BHT were used as positive control. e bacterial cultures were grown in the liquid medium at a suitable temperature. e bacterial strains were incubated on micrococcus, nutrient, and YM media and then cultured with shaking for 12 h at 37°C or 30°C. Two plant extracts, which were, respectively, collected within SFE and HD process, were screened for antimicrobial activity using a serial twofold dilution assay [15,16]. e minimum inhibitory concentration (MIC) of each compound was defined as the lowest concentration that inhibited microorganism growth. Bacterial growth was evaluated visually based on the degree of turbidity. Citral was used as positive control.

Paper Disc Diffusion Assay.
e antimicrobial activities of the plant extracts and fractions were determined by paper disc diffusion assay [15,17,18]. e diameter of each inhibitory zone was measured (in mm). Negative controls were prepared using the same solvents employed to dissolve the plant extracts. Citral was used as positive control.

Analytical
Procedures. All samples were prepared and analyzed in triplicate. All figures were obtained by using one of Origin 8.5, Microsoft 2010, Design-Expert 8.0.6 and onimic 9.0. To verify the statistical significance of all parameters, the values of mean ± SD were calculated. e yield of essential oil obtained by HD and SFE were determined by gravimetric analysis. e extraction yield (%) of essential oil was calculated as follows: where Y 1 is the extraction yield (%); ρ is the destiny of the essential oil; andV 1 is the volume of the essential oil. e capability of scavenging the DPPH radical was calculated using the following equation: where A c is the absorbance of the bland; A i is the absorbance of the sample reacted with the DPPH; and A j is the absorbance of the sample. e scavenging ability of 1,2,3phentriol of essential oil to self-oxidize was calculated using the following equation: where Y 2 is the scavenging ability; S 1 is the slope of the sample; and S 2 is the slope of control.

Determining Levels for Independent Variables.
In recent reports, many researchers have indicated the efficiency of extraction would be impacted significantly by some factors in SFE processing, such as extraction time, extraction temperature, and CO 2 flow, which were respectively studied to evaluate the impact of parameters on the SFE; the results are shown in Figure 1. Firstly, the effect of different extraction time periods on the extraction essential oil yield was examined at 30, 60, 90, 120, and 150 min in the process of SFE. e essential oil produced during SFE with respect to extraction time on extraction yield is shown in Figure 1(a). Results in the extraction yield of essential oil was increased with the extraction time going up from 30 min to 150 min. When the extraction time was shorter than 60 min, the oil was extracted significantly from the C. citronella but the yield was still at a lower level. At 120 min, the extraction yield of essential oil reached 4.0% (w/w on origin feedstock). ereafter, the extraction yield was almost stabilizing with the extension of time, which would increase the cost of SFE treatment. erefore, it is favorable to choose the range of 110-130 min as a desirable holding time for the SFE of essential oil.
Extraction temperature was one of the factors affecting extraction effectiveness. According to the above results, the influence of extraction temperature at 20, 25, 30, 35, and 40°C was studied (Figure 1(b)). As seen in Figure 1(b), the extraction yield of C. citronella essential oil reduced significantly with the extraction temperature going up from 20°C to 40°C. us, there are obviously a large number of the essential oils extracted during the SFE, from 2.6% to 4.0%. e reason maybe the viscosity of supercritical fluid decreased and the diffusion coefficient of fluid increased with the elevation of extraction temperature, meanwhile the Brownian motion of essential oil in C. citronella was enhanced so that accelerated the processing of mass transfer, resulting in improving the dissolubility of oil [2]. At 35°C, for example, the extraction yield of essential oil increased to 4.2% (w/w in original feedstock), approximately. However, no obvious variation in extraction yield was shown after 35°C. In addition, the constituent and structure of essential oil would be degraded and destructed according to some reports [8]. Considering the economical and extraction efficiency, 32°C-38°C was chosen as a desirable holding temperature during SFE process. e influence of CO 2 flow of 8, 12, 14, 18, and 22 L/h is shown in Figure 1(c), which indicated that the flow of CO 2 was a notable factor influencing the extraction results. When the CO 2 flow increased from 8 L/h to 16 L/h, the extraction yield of essential oil increased sharply, achieving from 2.4% to 3.6%. With the increasing flow of CO 2 , nevertheless, there was an abnormal decrease in extraction yield of essential oil. Some researchers indicated that it would be conducive to the area of effective contact between feedstock and CO 2 at the lower velocity condition so that it could be much easier to extract the essential oil from original [19]. Besides, when the flowing of CO 2 was too higher, it would possibly lead to the inadequacy for the area of contact, but also part of extracted oil could be brought back to the extractor from the separator, in which the mixture was not separated thoroughly for the shorter separation time. It was, thus, of interest to obtain essential oil within limits of CO 2 flow. e effects of extraction pressure on the extraction yield of essential oil were studied with five levels at intervals of 5 MPa from 10 to 30 MPa. As illustrated in Figure 1(d), with an increase in the extraction pressure from 10 to 25 MPa, the yield of essential oil was significantly increased (from 2.7% to 4.0%). As the extraction pressure rose from 25 to 30 MPa, the extraction yield of essential oil changed little (∼4%). is was probably, according to the supercritical extraction principle reported by some researchers [2,7,19,20], due to the enhancement in density and diffusivity of supercritical fluid with the increasing extraction pressure, which prompted the improvement of the strength of extraction. However, it was found that higher extraction pressure decreased the mass transfer time, even carrying part of extracted essential oil in a separator. Consequently, the extraction yield would be stabilized within higher extraction pressure. In consideration of cost, within the range of 22-28 MPa, was the appropriate extraction pressure during SFE processing in this study.
By comparison, some recent studies had reported that various factors existed in the separation reactor would have had an influence on the yield of active substance extracted from biomass during SFE processing [21]. Accordingly, some key factors, such as separation pressure and separation temperature, were evaluated for the SFE treatment in this part. As shown in Figures 1(e) and 1(f), the effect of separation pressure and separation temperature was investigated from 2.5 MPa to 4.5 MPa and 20°C to 40°C, respectively. It was abnormally apparent that the separation pressure and separation temperature resulted in an insignificant impact on extraction yield, wandering in 3.5% approximately. e result was beyond the expectation and inconsistent with previous results that the subsequent separation process would play a restrictive role on extracting essential oil, indicating the supercritical CO 2 had an outstanding capability of efficient extraction of essential oil with SFE, which was not affected by separation pressure and separation temperature. Consequently, in consideration of cost, 3 MPa and 25°C (at room temperature) were selected to be the separation pressure and separation temperature in the next experiment.

Construction of the SFE Processing.
After the singlefactor study about the evaluation of the impact of various parameters on the SFE treatment, as shown in Figure 1, the results had revealed the various components that changed tendency in the material, including extraction time, extraction temperature, CO 2 flow, extraction pressure, separation pressure, and separation temperature, respectively. ereinto, the separation pressure and separation temperature were not considered in this experiment. Although a large number of essential oils had been extracted from C. citronella in the SFE treatment, the high efficiency needs severe treatment conditions that would increase the cost of SFE treatment. In addition, it would be also increase the cost because of interaction of various variables. Accordingly, these variables were taken into careful consideration by using Box-Behnken design (BBD) with the soft of Design-Expert V8.0.6 in this study. e levels of each factor are given in Table 1. Twentyseven experiments were carried out from the design ( Table 2). e results of the second-order response surface model fitting for extraction yield in the form of analysis of variance (ANOVA) are given in Table 3. To test the fit of the model, the regression equation and determination coefficient R 2 were evaluated. e model presented a high determination coefficient (R 2 � 0.976) explaining 97.6% of the variability in the response (Table 3). e coefficients of regression were calculated and the following regression equation (4) was obtained. Table 3 presents Fisher's F-test of ANOVA, which also proves that this regression model was highly statistically significant (P < 0.0001). At the same time, the relatively lower value of the coefficient of variation (CV � 3.25 < 10%) indicates good precision and reliability of the experiments carried out. Adequate precision for our model has a signalto-noise ratio of 18.571 (>4) which indicates an adequate signal. e signification of each coefficient was determined by Tand P which is also listed in Table 3, which demonstrates that all the linear model terms (X 3 and X 4 ) and quadratic model term (X 1 X 2 , X 2 X 4 , and X 3 X 4 ) were significant. Moreover, the interactive model terms, X 1 X 2 , X 2 X 4 , and X 3 X 4 , were significant (P < 0.05), while the interactive model terms of X 1 X 3 , X 1 X 4 , and X 2 X 3 were insignificant (P > 0.1).

Journal of Analytical Methods in Chemistry
In addition, the empirical model (Eq. (1)) proved that the predicted data of the response from the empirical model are in agreement with the observed ones in the range of the operating variables. On the other hand, the value of adjusted (R 2 � 0.948) is also very high that indicated a high significance of the model. e maximum extraction yield obtained by using above selected variables was 4.36%, and the experimental maximum obtained was 4.32 ± 0.06%. e data show that predicted data on the response from empirical model were in agreement with those observed in the range of the operating variables. e coefficients of regression were calculated, and the following regression equation was obtained: where Y � response (extraction yield), X 1 � extraction time, X 2 � extraction pressure, X 3 � extraction temperature, and X 4 � CO 2 flow in coded values.
In order to gain a better understanding of the effects of the variables on the essential oil of extraction yield from C. citronella leaves, the predicted model was presented as 3D/ 2D response surface graphs, as shown in Figure 2, respectively, which were generated with one variable kept at its optimum level and varying the others within the experimental range (shown in Table 1). ese variables were optimized as follows: extraction time 120 min, extraction pressure 25 Mpa, extraction temperature 35°C, and CO 2 flow 18 L/h for the SFE process. Under optimized condition, the extracted essential oil reached 4.40%. Nevertheless, the result is comparable to that reported elsewhere on extraction of essential oil with various extraction methods, indicating that the SFE had significantly higher extraction efficiency and was the most energy-saving method than other methods, combining with an auxiliary instrument that could raise the cost of production [1,5,[22][23][24]. Consequently, these results have suggested that the SFE treatment is effective for selectively extracting the essential oil from C. citronella and may be expected to use in others active components extracted from various plants in future.
Recently, many researchers argued the extraction yield and essential oil compositions, but paid little attention to the effect of extraction methods on the essential oil compositions. It would limit the development and application of essential oils [5]. Accordingly, in the next experiment, constitute features of the extracted oil are characterized with modern analytical equipment.

Analysis of the Extracted Essential Oil.
e chemical compositions identified in the essential oils extracted by HD and SFE are summarized in Table 4. ere were 35 compounds identified in HD essential oils while 41 in SFE essential oils. ese compositions were grouped into seven classes (hydrocarbons, alcohols, aldehydes, esters, ketones, alkanes, and others) according to their functional groups. Alcohols and aldehydes were the main compositions in the essential oils extracted by HD and SFE. e abundant alcohols can promote the C. citronella leaf essential oils used for cosmetic applications. e contents of flavourless alkanes in SFE oil were much lower than those of SFE (Table 4), indicating that the essential oil in SFE contains fewer impurities, which were similar to the previous report [8]. e main component of essential oil extracted by SFE and HD was geranialdehyde and the next was geraniol, which were consistent with previous studies [23,25]. Furthermore, the content of this major compound in SFE extracts was higher than HD. is proved that supercritical fluid did not alter the main effective components in C. citronella essential oil, but the content of effective composition and extraction yield of essential oil was higher [8]. Journal of Analytical Methods in Chemistry 3.4. Antioxidant Activity Analysis. In this study, antioxidant activity of essential oil from the C. citronella leaves was evaluated using DPPH scavenging and 1,2,3-phentriol selfoxidation activity assays. e IC 50 values of essential oil are summarized in Table 5. As reports, 1,2,3-phentriol rapidly autoxidizes in the autoxidized solution, and several intermediate products, such as O 2 − , were formed, which resulted in the solution to become yellow brown with a spectrum showing a shoulder between 400 and 425 nm. e antioxidants could interfere with 1,2,3-phentriol autoxidation by acting as scavengers of O 2 − [11,12]. erefore, antioxidant ability can be determined by the scavenging activity of selfoxidation of 1,2,3-phentriol. Table 5 shows that both essential oils, extracted by SDE and SFE, respectively, had much more effective scavenging power for self-oxidation of 1,2,3-phentriol than BHTand ascorbic acid and it should be explored as a potential antioxidant with SFE process. Besides, the model of scavenging the stable DPPH radical is a widely used method of evaluating the free-radical scavenging ability of natural compounds. In the DPPH test, the antioxidants were able to reduce the stable DPPH radical to the yellow-colored diphenylpicrylhydrazine. e effect of antioxidants on DPPH radical scavenging was thought to be due to their hydrogen-donating ability [12].
e DPPH scavenging activity of the essential oils expressed in terms of IC 50 is shown in Table 5, with the higher antioxidant potency of DPPH scavenging compared with the antioxidant potency of BHTand ascorbic acid. ereinto, the essential oil extracted by SFE had the highest antioxidant potency of DPPH scavenging. e antioxidant activities of antioxidants have been attributed to various mechanisms, and one test is normally not enough to evaluate precisely the antioxidant activity of potential antioxidant [11]. erefore, we used these two assays to evaluate the total antioxidant capacity of essential oil from the C. citronella leaves. e results indicated that the essential oil had a noticeable effect on scavenging self-oxidation of 1,2,3phentriol and were able to scavenge free radicals.

Antimicrobial Activity Analysis.
e antimicrobial activities of extracts gained from HD and SFE had been   Journal of Analytical Methods in Chemistry assessed using a serial twofold dilution assay and paper disc diffusion assay, according to Hu et al. [23,24,26]. Table 6 shows the variation in the antimicrobial properties of the essential oil extracted from the C. citronella leaves. Among bacterial strains, the E. coli has the strongest inhibitory effect by the essential oil. Meanwhile, the essential oil shows the antifungal or anticandidal activity as well. On the other hand, more precise data concerning the antimicrobial properties of the extracts were obtained by paper disc diffusion assay [26]. Table 6 indicates the maximal inhibitory zones for each of the microorganisms that were sensitive to the extracts, which were in the range of 15-35 mm. ereinto, the greater inhibition was observed in the case of the SFE extracts than the others. Hence, the SFE extracts was significant antimicrobial or antifungal activities against all of the microorganisms tested, except A. niger.
In summary, the SFE extracts have significant antimicrobial activity against E. coli and A. niger, respectively. Antimicrobial activity in essential oil depended on mostly aldehydes and acid esters [22,27]. Recently, some literature reported there was the highest level of activity against E. coli with tea oil, tannic acid, and saponins [6,20,23]. Accordingly, these observations suggested that the antimicrobial activity of essential oil may be due to the presence of ester bond and ether bond [28]. Other phenolic acid-like phenols were also thought to contribute to plant defenses against pests and pathogens [11].

Conclusions
e essential oil was attained from C. citronella, a common herbaceous plant, by supercritical fluid extraction, resulting in 4.40% of essential oil extracted under mild condition. Alcohols and aldehydes were the main compositions in the essential oils extracted by HD and SFE. Compared with HD essential oils, the SFE essential oils contain fewer impurities and had better antioxidant and antimicrobial activities. Supercritical fluid extraction has presented an outstanding feature in efficient extraction of active components of the plant, which can be a desirable candidate for the current extraction processing.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.