Extraction and characterization of essential oil components based on geraniol and citronellol from Java citronella (Cymbopogon winterianus Jowitt)
- First Online:
- Cite this article as:
- Wany, A., Kumar, A., Nallapeta, S. et al. Plant Growth Regul (2014) 73: 133. doi:10.1007/s10725-013-9875-7
- 691 Views
Citronella oil is the main product of Java citronella grass (Cymbopogon winterianus Jowitt) rich in geraniol and citronellol, widely used in mosquito repellents and perfumeries. The age of the plant plays a key role in oil composition and its yield such that young leaves have lesser oil content than the mature leaves. Also, a remarkable difference between fresh and dried leaves regarding oil yield is observed. The various methods of extracting essential oils from citronella grass with respect to yield (%) were studied. Average percent yield in the manual extraction and hydro-distillation procedure was 0.8 and 1 % respectively, which was better as compared to steam distilled oil (0.7 %). The chromatographic analysis of essential oils with respect to standards geraniol and citronellol were studied by high performance thin layer chromatography (HPTLC) with n-hexane and ethyl acetate (3:2) as mobile phase followed by its separation on plates. The developed plates showed geraniol, citronellol and citronellal as major bands. The analysis of all extracted oil samples by means of electrospray ionization-mass spectrometry (ESI-MS) in the positive ion mode showed rapid mass fingerprints of constituents present in the samples according to the observed mass of standards. Furthermore, the analysis of vibrational spectra was accomplished with Fourier transform infra-red spectroscopy (FTIR) specifying all the functional groups as major peaks confirming all of them as monoterpene alcohols with conjugated double bonds. Thus, HPTLC, ESI-MS and FTIR studies evidenced that the two essential oil components were majorly present in the methanol extract suggesting methanol as a good extractant in the manual extraction process.
KeywordsCitronellolElectrospray ionization-mass spectrometryFourier transform infra-red spectroscopyGeraniolHigh performance thin layer chromatography
Cymbopogons are highly stress-tolerant plants that survive diverse edapho-climatic conditions prevailing in tropical and sub-tropical areas of Asia, Africa and America (Sangwan et al. 2001; Shasany et al. 2000). Citronella (Cymbopogon winterianus Jowitt) is an aromatic grass belonging to the Poaceae family, which gives essential oils upon steam distillation. Essential oils are concentrated essences extracted from different parts of plants, containing many substances, but little of them really characterize the fragrance (Mendes et al. 2007). Oil of citronella grass is one of the important essential oils obtained from different species of aromatic grasses. The onset of monsoon is the best time for the plantation of Java citronella but, can be initiated any time during the year.
Industrial interest in essential oils is due to their application as fragrances in perfumes, as flavour additives for use in food products or even as pharmaceutical products and in mosquito repellent manufacturing throughout the world. C. winterianus essential oil has CNS effect and anticonvulsant properties (Blanco et al. 2007). Citronella oil has been used as flavouring for foods and beverages in very low quantities [(approximately 45 ppm), Katz et al. 2008]. In traditional medicine, the oil has been used as an aromatic tea, vermifuge, diuretic, and antispasmodic. Perhaps the most widely recognized use for the oil is as an insect repellent (Simic et al. 2008; Silva et al. 2011).
Commercially, citronella oil is classified into two chemotypes, Ceylon citronella oil, obtained from Cymbopogon nardus (inferior type), and Java type citronella oil obtained from Cymbopogon winterianus (superior type). The Ceylon chemotype consists of geraniol (18–20 %), limonene (9–11 %), methyl isoeugenol (7–11 %), citronellol (6–8 %), and citronellal (5–15 %). Java chemotype consists of citronellal (32–45 %), geraniol (11–13 %), geranyl acetate (3–8 %), and limonene (1–4 %) (Katiyar et al. 2011). The two cultivated types are distinguished by the shape and length of their leaves. The new cultivars from the existing varieties of citronella have been developed through new breeding techniques viz. open seed pollinating systems (OSPs) and crop improvement programs. C. flexuosus, C. pendulus, C. winterianus and C. martini var. motia—are commercially cultivated as modern cash crops for essential oil production (Padalia et al. 2011). C. winterianus essential oil is rich in monoterpenes alcohols such as citronellal, geraniol, nerol and citronellol (Cassel and Vargas 2006; Katiyar et al. 2011).
The extraction of oil from citronella has been done by different processes such as fractional distillation, steam distillation and hydro-distillation (Clevenger’s apparatus) and through supercritical fluid equipments. These extracts are then subjected to GC analysis coupled with MS or other advanced analytical techniques to quantify the individual constituents present in the extracts (Khanuja et al. 2005; Vargas et al. 2010; Beneti et al. 2011; Silva et al. 2011; Setiawati et al. 2011). Another technique, which qualitatively analyses the oil samples, is TLC, but a more advanced, reliable, improved, quantifiable and a higher resolution with the ability to check purity of the samples is given by high performance thin layer chromatography (HPTLC) (Lehri et al. 2011). When many of the targeted metabolites are not readily amenable to GC/MS analysis due to volatility issues or when searching for unknown metabolites, LC/MS is best for this type of discovery based approach. Electrospray ionization (ESI) is most commonly used ionization technique in LC/MS. However, a valuable alternative to thermospray as an interface between HPLC and mass spectrometry can be ESI, which permits detection of low concentrations of analytes. Therefore, ESI-MS approach is more advantageous because it relies on its soft ionization characteristics, an important prerequisite in the case of thermolabile compounds (Catharino et al. 2005). Also, if the compound is unknown and is well identified by HPTLC, analysed by ESI-MS then, functional group analysis can be done with the help of Fourier transform infra-red spectroscopy (FTIR). It is an excellent tool for qualitative and quantitative analysis of compounds (Schulz et al. 2005). A combination of these three techniques can be very useful to identify an unknown component present in an oil sample, purify it and resolved into its ions.
The present investigation deals with the different extraction methods such as manual extraction for geraniol and citronellol using various solvents as well as steam distillation and hydro-distillation. Finally, the presence of geraniol and citronellol were characterized by HPTLC, ESI-MS and FTIR from Java citronella.
Materials and methods
Collection of Plant materials, standards and other chemicals
Leaves of Cymbopogon winterianus Jowitt (Java citronella) were collected from the Pharmaceutical Indigenous Medicinal Plant Research Farm, BIT, Mesra. A flowering specimen (herbarium) was prepared and sent to the Botanical Survey of India, Kolkata and was also botanically identified by Dr. S. Jha, Department of Pharmaceutical Sciences, BIT, Mesra, Ranchi. The leaves were shade dried at 29 °C (RT) until the leaves became brittle. The dried leaves were kept for steam distillation. The fresh leaves (350 g) were also collected, cut into small pieces for hydro-distillation and part of them was powdered using liquid nitrogen for preparation of crude extracts using different organic solvents. The essential oils standards were obtained from Sigma-Aldrich USA. The chromatographic plates (HPTLC) and the analytical grade solvents were from Merck (Germany).
Extraction of essential oils
Preparation of crude extract
Dry powder (25 g) of citronella leaves was macerated with 100 ml each of methanol, hexane and petroleum ether solvents. The macerates were kept for incubation at 45 °C for 2–3 h in a rotary shaker at 110 rpm. The extract was filtered through layers of filter paper and the green colored filtrate was collected in a fresh conical flask. Chlorophyll was removed by passing this filtrate through activated charcoal, two–three times until a pale coloured filtrate is obtained. A clear extract was obtained when filtered through anhydrous sodium sulphate to remove the water content. Further, the extract was evaporated by heating in the heating mantle at 45–50 °C till the oil droplets appeared. A minimum quantity (500 μl) of the respective organic solvents were added and stored in brown vials at RT for further phytochemical analysis. The extracts obtained were named as ME (methanol extract), HE (hexane extract) and PE (petroleum ether extract).
Extraction of oil through steam distillation
Extraction of oil through hydro-distillation
TLC and HPTLC analysis
The automated HPTLC (HPTLC pre-coated plates silica gel 60 F 254; Plate size—6.0 × 10.0 cm) system was equipped with CAMAG TLC Scanner 3 supported by winCATS Planar Chromatography Manager software (v 1.3.3). Ready-to-use silica coated plates were activated by keeping them at hot air oven for 5 min (110 °C) and placed in the automatic sample applicator. The HPTLC was programmed to automatically spray 3 μl of each sample and standard in band form (4 mm band-width) using specialized Linomat injector device on one-side of the TLC plate in individual tracks. The plate was developed in the developing chamber (CAMAG) containing the developing solvent until the solvent front reached the maximum distance. The developed plate was dried with a plate drier and subjected to UV analysis (wavelength range 200–800 nm) in the CAMAG TLC Scanner 3. Tentative identification of spots was achieved by comparison of Rf values with those of reference standards after complete run.
The developed plates were initially examined by iodine vapours and thereafter visualized by spraying with vanillin reagent in 5 % ethanolic concentrated sulphuric acid till grayish-purple bands appeared and then heated at 100 °C for ten minutes. Appearance of blue, violet and pink spots was considered as positive for monoterpenes alcohols (Geraniol and citronellol). Also, the plates were examined under UV light and photographed.
Electrospray ionization-mass spectrometry fingerprints of oil extracts were obtained in the positive ion (ESI+) mode on a MicrOTOF-Q II MS system (Bruker Daltonics India Pvt. Ltd, Bangalore). The default ESI-MS conditions were followed. ESI-MS/MS of selected ions were acquired by low-energy (15–30 V) collision-induced dissociation. Prior analysis, the oil samples (250 μl) were homogenized in a flask with a solution containing equal parts of water and acetonitrile, completing the final volume of 1.0 ml. The phases were allowed to separate, and the top (hydro-alcoholic) layer was removed. 0.1 % Formic acid (10 μl) was added to each sample for ESI (+)-MS analysis. These solutions were then injected at a flow rate of 10 μl min−1 using a syringe pump, and mass spectra were acquired over the 50–1,500 m/z range.
The vibrational spectra were recorded in the range between 3,900 and 450 cm−1 with an ATR/FT-IR spectrometer “IR-Prestige 21” (Shimadzu Corporation, Japan) in a diffuse reflection configuration. The instrument was fitted with a Michelson interferometer with germanium coated KBr plates capable of analysing solid, liquid and gaseous samples. 5–10 μl of the essential oil sample were placed on the surface of the ATR crystal (diameter 0.5 mm2). It evaluates the vibrational spectra with respect to % transmittance (%T) and wave-number (7,500–350 cm−1).
The plant was collected from the medicinal plant research farm of our Institute. The botanical identity of the plant specimen was identified by Dr. S. Jha, Professor, Department of Pharmaceutical Sciences, BIT, Mesra. Also, the flowering specimen was submitted to the Botanical Survey of India, Kolkata and was confirmed as Cymbopogon winterianus Jowitt (Specimen no-12235). Various combinations of solvents for extraction and developing the chromatograms were attempted during the isolation of active compounds from the plant.
Extraction of essential oils
Three methods of essential oil extraction from Java citronella have been studied. The first method deals with the manual and crude extraction of essential oils using three different organic solvents (S1: methanol, S2: n-hexane, S3: petroleum ether) from dry powder (25 g) of citronella leaves. The second method involves extraction in a conventional steam distillation plant, which is operated in the Institute’s research farm for experimental purpose. The third is the most suitable and relatively advanced method of extraction using Clevenger’s apparatus for hydro-distillation. Average percent yield in the manual extraction and hydro-distillation procedure was 0.8 and 1 % respectively. But the average percent yield in steam distillation was 0.7 %, which is still less than the other two. Moreover, the time required for the extraction purpose in hydro-distillation (>8–10 h) and steam distillation is more as compared to manual extraction, which takes 5 h (approx.). Methanol and n-hexane among the three organic solvents could be considered as manual extractants. But, the average percent yield of oil was 1 % in methanol was much high in comparison to n-hexane (0.6 %), again due to its volatile nature thereby proving methanol as a good extractant.
TLC and HPTLC analysis of extracted oils
Rf values of standards and samples obtained in HPTLC as well as in TLC plates
Rf values (HPTLC)
Rf values (TLC)
Steam distilled oil
0.70, 0.85, 0.92
0.68, 0.80, 0.87
Hydro distilled oil
0.70, 0.85, 0.94
0.68, 0.79, 0.86
0.65, 0.70, 0.80 (faint)
Expected components present in the sample w.r.t. their m/z values observed in ESI-MS fingerprints
Retention time (mins)
Observed mass (m/z)
Analytical evaluation of the FT-IR spectrum showing major functional groups
Group frequency wave number (cm−1)
Steam distilled oil
Functional group assigned
Polymeric –OH stretch
Methylene C–H asymmetric stretch
Terminal aldehydic C–H stretch
May be aldehyde
Olefinic unsaturation C=C
Methylene C–H bend
Methyl C–H symmetric bend
Gem-dimethyl or “iso”-doublet
C–C stretch skeletal vibrations
May aryl-O–H stretch
Simple –OH stretch
CH=CH trans unsaturation
1, 3 disubstitution/1, 4 disubstitution
A number of methods for the extraction of plant metabolites from several medicinal plants have been reported. But, these methods are not universally applied for all medicinal plants. The plant, Cymbopogon winterianus for the study was selected because of very less information towards the characterization of its metabolites.
Extraction methods and yield of oil
All the three methods incorporated in the study gave a considerable amount of essential oil. In the manual extraction method, the advantage was the fewer amounts of the sample (25 g; dry leaves powder) with respect to steam distillation (1,000 kg; dry leaves) and hydro-distillation (300 g; fresh leaves). The average percent yield obtained in the manual extraction and hydro-distillation procedure was quite comparable and is in accordance with the yield obtained by Silva et al. (2011). But, the average percent yield of essential oil extracted from the steam distillation plant, which is still less comparable with respect to the other two, due to the large size of the steam distillation plant, continuous heat loss from the plant and continuous operation of the plant for 14–16 h with huge amount of raw material. This relative percent yield of essential oil using steam distillation is close to the result obtained by Cassel and Vargas (2006) who obtained 1.02 % yield in 12 h of distillation. However, the time required for the extraction purpose in hydro-distillation and steam distillation is more because of the complexity of the apparatus and the specifications of the equipment as compared to manual extraction, which is performed in relatively easy steps utilizing no major equipments except that the leaves should be in powdered form. Beneti et al. (2011) used vacuum fractional distillation for the extraction of essential oil from C. winterianus, Silva et al. (2011) and Setiawati et al. (2011) extracted the oil through hydro-distillation (Clevenger’s apparatus) and supercritical fluid extraction methods respectively whereas Cassel and Vargas (2006), Jirovetz et al. (2006) and Vargas et al. (2010) used the steam distillation procedures, which implies that simple procedures like manual extraction has not been attempted till date. With respect to these literatures, present study though incorporated the extraction procedures based on hydro-distillation as well as steam distillation but confined mainly on manual extraction. Thus, this crude extraction method using the three organic solvents has been reported for the first time in this study. Among the three organic solvents, methanol and n-hexane proved best for the extracting oils but petroleum ether being highly volatile could not be considered as an extracting solvent (extractant) manually. Hence, it may be inferred that manual extraction can be performed when no distillation apparatus is available in the laboratory. The leaves can be powdered using an electrical grinder so that the oil extraction may get facilitated.
TLC and HPTLC analysis
A number of different combinations of mobile phases were checked in trial separations TLC for optimizing the developing solvent so that it can be applied to HPTLC system. The standards and samples, which were diluted in methanol gave comet shaped spots and no resolution of samples was observed. Conversely the samples and the standards resolved distinctly when diluted in n-hexane and gave comparable Rf values, which were in agreement with the result observed by Nigam et al. (1965), who performed the same experiment with TLC silica gel G plates and not in HPTLC plates. The ME and HE samples yet diluted in n-hexane were not resolved in HPTLC due to the micro size of injector and the presence of impurities in the extracted samples. The same set up was checked in conventional TLC silica gel G plates. Despite the three spots in HO and SO merged but still two of them are visible. The ME and HE gave three and two spots respectively but after some time vanished (Fig. 3). The samples (SO and HO) also gave the same Rf as the standards showing three peaks as they are crude in nature and will consist of other essential oil components. The HPTLC profiles of SO and HO were quite same.
In routine TLC experimental, the detection is only by spray method and the Rf value is not accurately recorded. However, UV based scanning after developing HPTLC plate not only provides opportunity for scanning at specific wavelengths but is also useful for quantification. No reports have been evidenced for the HPTLC quantification of geraniol and citronellol from citronella oil. Rather, HPTLC densitometric method of analyzing geraniol in palmarosa oil from rosha grass (Cymbopogon martini) has been studied (Lehri et al. 2011). Thus, this preliminary but fundamental study on HPTLC analysis of citronella oil obtained from citronella grass (Cymbopogon winterianus) could be beneficial to all the researchers.
ESI-MS analysis of extracted oil
The ESI-MS analysis of standards and samples showed a number of peaks and in the 50–1,500 m/z range, each oil extract produced numerous ions allowed direct oil classification based mainly on much defined changes in relative cation abundances. These ions appear in the fingerprints (Fig. 8a–g) appeared to be characteristic for the essential oils. Methanol extract, which is a crude oil sample, showed the presence of most of the major ions in the initial m/z scan w.r.t. less Rt. However, the hexane extract gave no peak that corresponds to the two standards but large number of peaks was obtained, which indicated the presence of noise and impurity. The ESI-MS fingerprint of HO indicated only two of the major ions. Similarly, the steam distilled oil showed only one of them.
It is therefore, not necessary for fingerprinting classification, that ESI (+) MS will permit the tentative identification of some of the ions via comparing with reported data or structure with the related database. Thus, ESI-MS is a suitable technique able to provide fast fingerprint characterization of complex natural mixtures such as beer, whisky, wine, spices, gasoline, and crude oil (Catharino et al. 2005). Therefore, the MicrOTOF-Q II system provides exact mass accuracy and resolving power to expand the horizon for GC-MS based analyses.
It allows the qualitative determination of organic compounds with respect to its molecular group, which shows a characteristic vibrational mode with the appearance of bands in the infrared spectrum at a specific frequency influenced by the surrounding functional groups (Fig. 9a–e). In the attenuated total reflectance (ATR) IR spectra, the most intense peaks were at 3,400–3,200 cm−1, which attributed to the polymeric hydroxyl group (OH) stretch, which confirmed that the samples were in the vapour state or a non-polar liquid and is observed at the naturally high wave number. These are the most significant signals used for hydroxyl groups shown by the samples and the standards. Another intense peak was in the range 2,935–2,915 cm−1, which contributed to the methylene C–H asymmetric stretch that are saturated aliphatic alkyl group wave number range present in all the samples and the standards. The terminal aldehydic (C–H) moiety has been shown by SO and HO in the range 2,800–2,700 cm−1 contributing to the carbonyl compound group range. Also, SO has shown an aldehyde group (1,740–1,725 cm−1) but a ketone (1,725–1,705 cm−1) may be present in the HO and the ME. The olefinic unsaturation C=C stretch or the conjugated double bond system showed a relatively narrow, strong absorption above 1,650 cm−1, which confirms the presence of conjugated double bonds in geraniol and citronellol. Methylene C–H bend and methyl C–H symmetric bend was prominent in 1,485–1,445 and 1,380–1,371 cm−1 respectively in the SO and HO samples. An intense –C–O– stretch is seen around 1,140–1,050 cm−1 in all the samples and standards confirming a terminal hydroxyl group in the structure. Minor vibrations of C–C skeleton, C=C skeleton, simple –OH stretches, CH=CH trans unsaturation and 1, 3 disubstitution/1, 4 disubstitution was also observed around 1,350–1,000, 1,680–1,620, 910–860 and 860–800 cm−1 respectively.
The overall spectra confirmed the presence of geraniol and citronellol and other essential oil components in the extracted oil samples, which are in agreement with Schulz et al. (2005) and interpreted with the help of FTIR laboratory manual by Coates (2000). The methanol extract also showed promising results in terms of the presence of all the components which proves the success of the extraction method.
The present investigation was majorly confined to the manual, crude extraction of essential oils using organic solvents in easy steps, which can be performed in the laboratory and its analysis by the fundamental techniques has been reported for the first time in this study. Methanol being a polar extractant has the ability to extract essential oils with a moderately good yield (0.8 %), which proved good in this study. Furthermore, steam distilled oil (0.7 %) and HO (1 %) gave similar results when compared w.r.t. yield %, and essential oil components.
The strategy was applied such that if the components in the oil sample are unknown and can be well identified by HPTLC, analysed by ESI-MS then, functional group analysis can be done with the help of FTIR. A combination of these three techniques can be an excellent tool for qualitative and quantitative analysis of compounds and very useful to identify an unknown component present in an oil sample, purify it and resolved into its ions. The results obtained in HPTLC confirmed the presence of essential oil components in the manual extracts, steam distilled extracts and hydro-distilled extracts of citronella. ESI-MS is a suitable technique provides exact mass accuracy and resolving power to expand the horizon for GC-MS based analyses. It has the ability to provide fast fingerprint characterization of complex natural mixtures such as crude oils. The overall FTIR spectra confirmed the presence of geraniol and citronellol and other essential oil components in the extracted oil samples. Thus, it can be concluded that the methanol extract showed very considerable and promising results in terms of the presence of all the components, which proves the success of the extraction method and use of methanol as an extractant.
This preliminary yet fundamental study on essential oil extraction, analysis of citronella oil obtained from citronella grass (Cymbopogon winterianus) could be beneficial to researchers, pharmaceutical industries and perfumeries with respect to the amount of geraniol and citronellol present in the extracted oil samples manually and the purity of oil.
The authors sincerely thank Dr. S. Jha (Professor, Department of Pharmaceutical Sciences, BIT, Mesra, Ranchi, Jharkhand, India) for identifying the flowering specimen. UGC, New Delhi, India is gratefully acknowledged for providing financial support to Ms. Aakanksha Wany (F1-17.1/2012-13/MANF-2012-13-CHR-CHH-9539) as JRF. The authors are grateful for the facility of FT-IR provided by the Central Instrumentation Facility, TEQIP, Department of Biotechnology, BIT, Mesra. The technical support and help provided by Mr. Puspendra Patel (Research Scholar), Department of Pharmaceutical Sciences, BIT, Mesra for HPTLC and Clevenger’s apparatus is greatly acknowledged.