Separation and Purification Technology 133 (2014) 443–451

Contents lists available at ScienceDirect

Separation and Purification Technology

journal homepage: www.elsevier.com/locate/seppur

Supercritical CO2 extraction of Eucalyptus leaves oil and comparison

with Soxhlet extraction and hydro-distillation methods
Suwei Zhao, Dongke Zhang ⇑
Centre for Energy (M473), The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Oils were extracted from the leaves of Eucalyptus loxophleba ssp. lissophloia, also known as oil Mallee,
Received 2 August 2013 using a laboratory scale supercritical fluid extraction (SFE) system using CO2. The effect of temperature
Received in revised form 1 July 2014 (40–80 °C), pressure (10–50 MPa) and extraction time (30–150 min) on the oil yield was investigated
Accepted 3 July 2014
using a central composite design method to determine the significance and interactions of these param-
Available online 18 July 2014
eters. The results showed that pressure had the most significant enhancing effect on the oil yield, while
temperature and time showed a lesser impact. There was also pronounced interaction between temper-
Keywords:
ature and pressure and their combined effect on the yield was such that the oil yield increased with
Carbon dioxide
Eucalyptus leaves oil
increasing temperature at high pressures but decreased at low pressures. For comparison, the Eucalyptus
Hydro-distillation leaves were also extracted with Soxhlet extraction, using two different solvents, and hydro-distillation
Response surface methodology methods. In the Soxhlet extraction, the solvent type had a more significant effect on the oil yield than
Soxhlet extraction the extraction time. Overall, the Soxhlet extraction produced the highest oil yield while hydro-distillation
Supercritical fluid extraction the lowest among the three methods. The SFE yield was up to 4.78%, comparable in magnitude to that of
the hexane Soxhlet extraction of 7.9%. The chemical compositions of the extracted oils were analysed
using a gas chromatography-mass spectrometer (GC–MS). The dominant component identified in the
Eucalyptus oil was 1,8-cineole. The chemical compositions of the extracts were quite different for the
three extraction methods. The oil extracted by hydro-distillation contained only volatile compounds
while the oil from the SFE and Soxhlet contained both volatile and higher molecular weight compounds.
The effect of the addition of ethanol as a modifier at concentrations from 5 w% to 15 w% on the supercrit-
ical fluid extraction of Eucalyptus leaves oil was also investigated. The ethanol addition was shown to
increase the efficiency of oil extraction from Eucalyptus leaves and the oil yield increased with increasing
ethanol concentrations.
Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction
One of the plants with its various extracts that are extensively
used in the cosmetic, perfumery, food and pharmaceutical industry
is Eucalyptus [1]. Eucalyptus is a large genus of the Myrtaceae family
which includes over 700 species [2]. Although Eucalyptus is widely
grown in many countries all over the world, most of the species are
native to Australia. This plant is a genus of tall, evergreen and mag-
nificent trees cultivated all over the world for its oil, gum, pulp,
timber, medicine and aesthetic values [2]. Among the various
wood and non-wood products, Eucalyptus oil found in its foliage
is a most important one [2]. Eucalyptus oil is a complex mixture
of a variety of monoterpenes, sesquiterpenes, and aromatic phe-
nols,oxides, ethers, alcohols, esters, aldehydes and ketones. They
are extracted from the foliage of Eucalyptus trees, the quantity
⇑ Corresponding author. Tel.: +61 8 6488 7600; fax: +61 8 6488 7622.
E-mail address: Dongke.Zhang@uwa.edu.au (D. Zhang).
and strength of the oil vary across Eucalyptus species [2]. Eucalyptus
oil has been shown to contain very high amounts of 1,8-cineole,
which has chemical and physical properties that make it suitable
for a range of applications. The most-known compound, terpenoid,
gives Eucalyptus foliage its characteristic smell. Eucalyptus oil is
used widely as an ingredient in many general pharmaceutical
products (eg liniments, inhalants, expectorants) due to its broad
biological properties including anti-inflammatory, anti-allergenic,
anti-asthmatic, anticonvulsant, antiseptic, aquaculture antiviral,
anti-bacterial and anti-malarial. It has also been used as a flavour
and aroma enhancer in food and cleaning products and in cosmetic
formulations [3].
Eucalyptus oil can be extracted using a variety of methods. Cur-
rently, the most popular method of extraction is steam extraction
or hydro-distillation [4]. However, the high temperature operation
of this technique can lead to degradation of thermally labile
compounds and partial hydrolysis of water sensitive compounds,
resulting in the formation of artefact undesirable in the final

http://dx.doi.org/10.1016/j.seppur.2014.07.018
1383-5866/Ó 2014 Elsevier B.V. All rights reserved.
444 S. Zhao, D. Zhang / Separation and Purification Technology 133 (2014) 443–451
extracts [4]. Solvent extraction using ethanol has also been applied
[5], however, separating the solvent from the extracted is often too
difficult and some solvent residues would be present in the finished
product. Recently, supercritical fluid extraction (SFE) has gained
increasing attention particularly in the food, pharmaceutical and
perfume industries [6]. The commonly used fluid in SFE is CO2,
which has several unique characteristics and physic-chemical prop-
erties, being non-toxic, non-flammable, inexpensive, odourless, and
of low critical pressure (7.38 MPa) and temperature (31.1 °C) [6].
The use of CO2 leaves no residue in the products, thus providing
an oil of superior quality. Therefore, SFE using CO2 has several advan-
tages over traditional extraction techniques including operation at
low temperatures thus preservation of the thermally labile compo-
nents in the extracts. Besides, SFE can bring about environmental
benefits as it uses no or significantly less environmentally hostile
organic solvents. Although a few studies have explored the possible
use of supercritical CO2 (SC-CO2) extraction to extract Eucalyptus oil
[7,8], the conditions have not yet been optimized. In general, several
important factors, including pressure, temperature and extraction
time, have important effects on the oil yield.
The response surface methodology (RSM) has been demonstrated
to be a powerful tool for determining the factors and their interac-
tions [9]. The RSM is a collection of mathematical and statistical
techniques useful for analysis of problems in which a response of
interest is influenced by several variables and the objective is to opti-
mize this response [9]. This procedure involves fitting a function to
the experimental data and then using optimization techniques to
determine the optimum parameters [10]. It is much faster and more
efficient for gathering research results than the classic, one variable-
at-a time or full-factors experimentation approach [11].
There are several factors such as operating pressure, temperature,
solvent flow rate, extraction time and sample particle size that can
affect the performance of supercritical fluid extraction. The effect
of particle size on the supercritical fluid extraction of Moringa oleifera
seeds was studied by Zhao and Zhang [12]. It was found that the
extraction process was characterized by two periods. The first period
features a constant extraction rate which can be explained by the
extraction of solute more accessible to the solvent and the particle
size has little effect in this period. In the second period, the oil yield
increases with a decrease in particle size. This is because the intra-
particle diffusion has taken control of the oil transfer during this per-
iod. Besides, the effect of the solvent flow rate on the supercritical
fluid extraction of M. oleifera seeds oil was also investigated [13]. It
was found that in the case of M. oleifera seeds, the oil yield increased
with increasing the solvent flow rate at the same extraction time.
However, the solubility of M. oleifera seeds oil was not a function
of the CO2 flow rate but an indication of the mass transfer limitation.
The effects of sample particle size and solvent flow rate have been
well addressed in the literature and were not a primary objective
in the present work as described in this paper.
The objective of this study was to optimize the process param-
eters (pressure, temperature and extraction time) for the extrac-
tion of Eucalyptus oil using the response surface methodology,
and to evaluate the effect of these parameters on the oil yield. In
addition, the yield and chemical profiles of the extracts obtained
by SFE were also analysed and compared to those achieved by
hydro-distillation and Soxhlet extraction methods.
2. Materials, experimental design and analytical methods
2.1. Materials
Leave samples of a Eucalyptus loxophleba ssp. lissophloia, also
known locally as oil Mallee, were collected from the Narrogin
region, southwest of Western Australia. The samples were air dried
for two days and the final moisture content was determined, by
drying the samples in an oven (Model 8050, Contherm, New
Zealand) set at 103 °C for 5 h [14], to be 12.3%. The air dried sam-
ples were ground with a knife grinder (Model 3383-L30, Thomas
Scientific, USA) and the ground samples were sieved using a sieved
shaker (model EFL2000/2, Endecotts Ltd., London, England). The
fraction of particles under 400 lm was selected for all subsequent
extraction experiments. The final samples were kept in a sealed
container and placed in a refrigerator before experimentation.
2.2. Supercritical fluid extraction
The supercritical fluid extraction (SFE) experiments were car-
ried out using an SFT Custom SCW-SFE system (Newark, DE,
USA) as described elsewhere [12,15]. In this work, 5 g of Eucalyptus
leaves were carefully weighed and loaded into the 50 ml SC-CO2
extraction vessel. About 1 g of glass wool was packed at both ends
of the extractor to stop entrainment of the substrate. SFE started as
soon as the desired pressure and temperature had been reached.
The flow rate of the expanded gas CO2 (under atmosphere pressure
and room temperature of 20 °C) was set at 2 L/min in all runs,
determined based on a set of preliminary experimental trials
where the solvent flow rate was varied from 1.0 to 4.0 L/min and
the flow rate of 2 L/min was judged to be appropriate for this
study.
The extract was collected in ethanol in an amber bottle. In order
to improve the collection efficiency, the bottle was placed in an ice
bath during the dynamic extraction stage, which also acted as a
freezing-trap to minimise the loss of volatile compounds as the
sublimation of CO2 decreases the temperature of the collection sol-
vent. The precipitates in all the connection tube lines were washed
out with ethanol and then mixed with the collected extract in the
amber bottle. The mixture was made up to 10 ml with ethanol, and
1 ml was taken for GC–MS analysis, the rest was put in a rotary
evaporator (model N-1000S-W, EYELA, Tokyo, Japan) to remove
the ethanol solvent, and then the weight of extracts was measured.
Finally, the extraction yield was estimated as follows:
mass of total extracted
Extraction yield ð%Þ ¼  100% ð1Þ
mass of dried leav es
A three-factor central composite design (CCD) combined with
response surface methodology (RSM) was applied to determine
the best combination of process variables for SFE of the Eucalyptus
leaves oil. The independent variables studied here were pressure
(X1: 10–50 MPa), temperature (X2: 40–80 °C) and extraction time
(30–150 min), while the response variable was the oil yield. Table 1
shows the arrangement of the CCD performed in this investigation,
where 20 randomized experiments including six replicates at the
centre points were employed to fit the full quadratic equation
model. The yields presented in Table 1 were the average of at least
three measurements under otherwise the same conditions. The
polynomial equation employed in this study is shown as Eq. (2),
with which the linear (X1, X2, X3), quadratic (X21, X22, X23) and interac-
tive (X1X2, X1X3, X2X3) effects of independent variables of pressure
(X1), temperature (X2), and time (X3) on dependent variable (Y)
can be determined:
X
3 X
3 XX
3
Y ¼ b0 þ bi X i þ bii X 2i þ bij X i X j ð2Þ
i¼1 i¼1 i<j¼1
where Y is the response (extraction yield of Eucalyptus oil), b0 bi bii
bij are constant coefficients of intercept, linear, quadratic and
interaction terms, respectively. Xi and Xj are independent variables
(pressure, temperature and extraction time). The actual levels of the
independent variables used in the experimental design and
the observed response for Eucalyptus oil were shown in Table 1.
 S. Zhao, D. Zhang / Separation and Purification Technology 133 (2014) 443–451 445

Table 1 removed at 60 °C at about 10 kPa using a rotary evaporator (model

Central composite design for the supercritical fluid extraction of Eucalyptus leaves. N-1000S-W, EYELA, Tokyo, Japan). The extraction under each set of
Run # Factors Yield (%) conditions was performed in triplicate and the total extraction
X1 (pressure, MPa) X2 (temperature, °C) X3 (time, min) yield was obtained by the mean value of the extracted oil mass
divided by the mass of the raw material used, on dry weight base.
1 40 70 60 4.66
2 50 60 90 4.62
3 20 50 120 3.50 2.5. GC–MS analysis
4 30 60 90 4.04
5 30 60 30 2.73
The composition of the extracted materials was determined
6 30 60 150 4.39
7 10 60 90 0.99
using an Agilent 7890 N series gas chromatograph equipped with
8 20 50 60 2.99 an Agilent 5975 mass selective detector and a HP-5MS column
9 40 50 60 3.39 (30  0.25 mm (5%-Phenyl)-methylpolysiloxane column, film
10 30 60 90 3.96 thickness  0.25 lm). The GC settings were as follows: Oven tem-
11 20 70 120 3.06
perature was held at 40 °C for 5 min and then increased to 250 °C
12 30 80 90 4.44
13 30 60 90 4.29 at 2 °C /min and held at 250 °C for 30 min. Each sample (1 lL) was
14 20 70 60 2.51 injected into the column at a split ratio of 20:1. Helium was used as
15 30 60 90 4.18 the carrier gas at a flow rate of 1.0 ml/min. The temperature of the
16 30 60 90 4.26 ion source and transfer line was 230 °C and 250 °C, respectively.
17 40 50 120 4.45
18 40 70 120 4.78
The mass spectrometer was operated in the electron-impact ioni-
19 30 60 90 3.91 zation (EI) mode at an energy level of 70 eV. The scanning range
20 30 40 90 3.31 was 40–650 amu and the scanning rate was 0.2 s/scanning. The
percentage compositions of the identified compounds were calcu-
lated from the GC peak area without considering response factors.
The retention indices were determined relative to a homologous
The data collected from the SFE tests was analysed using the series of n-alkanes (C8–C40) under the same operating conditions.
‘‘Design expert’’ software (Design- Expert 7.1.3 Trial, State-Ease, The components of the oil were identified by comparing their mass
Inc., Minneapolis MN, USA). The statistical analysis of the model spectra against the NIST (National Institute of Standards and Tech-
was performed in the form of analysis of variance (ANOVA), values nologies) MS spectra library. The compounds were also confirmed
of p < 0.05 were regarded as significant. by comparing their retention indices with the data published in the
In order to investigate the effect of modifier addition on literature [8].
supercritical fluid extraction of Eucalyptus leaves oil, ethanol at
concentrations ranging from 5 w% to 15 w% were studied. The 3. Results and discussion
concentration of ethanol was expressed as the weight of ethanol
versus the weight of CO2 under the processing conditions. The 3.1. The effect of operating parameters in SFE
experiments were performed at 60 °C, CO2 flow rate of 2 L/min,
90 min extraction time at two different pressures of 10 MPa and The SFE experimental data was analysed using the analysis of
30 MPa, respectively. The ethanol contents remaining in the variance (ANOVA) to assess the ‘‘goodness of fit’’, which was listed
extracted oil samples were removed at 60 °C at reduced pressure in Table 2. As shown in this table, the probability (p) value of the
of 0.1 MPa using a rotary evaporator. quadratic equation was less than 0.0001 which implied that the
equation was an excellent fit and was very suitable for the present
2.3. Hydro-distillation experimental data. A statistically significant confidence level at
99.999% was obtained. A ‘‘lack of a fit’’ measures the failure of
The ground Eucalypts leave sample, prepared in the same way the equation to represent data in the experimental domain at
as those used in SFE, was also submitted to hydro-distillation in points which are not included in the regression. If the p-values of
a Clevenger-type apparatus. This simple Clevenger-type apparatus the ‘‘lack of fit’’ was smaller than 0.05, the model did not fit the
contains a 500 ml flask, a condenser and a measuring tube with data well [16]. The probability value of ‘‘lack of fit’’ was 0.0507,
stopcock, a return tube for the aqueous part of the distillate con- suggesting that the model was acceptable [16]. The coefficient of
nects the bottle of the measuring tube and the vertical tube. About determination (R2) obtained from the calculated quadratic equa-
15 g of a sample was immersed into 300 ml water and distilled for tion should also be taken into consideration when validating the
up to 5 h; the volatile distillate was collected over anhydrous model. The R2 value at 0.9537 was considered very good, indicating
sodium sulphate and kept in a refrigerator till GC–MS analysis. that this equation adequately represented the real relationship
Each of the experiments was repeated at least three times to among the parameters chosen. Furthermore, each experiment
ensure repeatability and statistical validity of the data. was repeated and the statistical analysis was performed so that
the experimental results can be used to valid the model with con-
fidence which is consistent with the report from the literature [17].
2.4. Soxhlet extraction The coefficients of the regression equation for the different sets
of the Eucalyptus oil yield data are shown in Table 2. Fitting the
Approximately 5 g of ground Eucalyptus leaves, prepared in the constants and coefficients into Eq. (2), the regression equation as
same way as those used in SFE, were also placed in a cellulose a function of pressure, temperature and time thus are obtained
thimble and transferred to a Soxhlet extractor. The extractor was as follows (in terms of coded levels):
filled with 150 ml of a solvent and heated under reflux for eight
hours through a water bath. Two solvents, hexane and ethanol, Y ¼ 4:12 þ 0:78X 1 þ 0:18X 2 þ 0:35X 3 þ 0:32X 1 X 2
were examined in the present work. The hexane extraction was þ 0:01X 1 X 3  0:11X 2 X 3  0:32X 21  0:05X 22  0:13X 23 ð3Þ
performed at water bath temperature of 78 °C for 1, 2 and 8 h,
while the ethanol extraction were performed at 90 °C for 0.5, 1, 2 where Y is the oil yield (%, dry weight of plant material), X1, the
and 8 h. After an extraction run was completed, the solvent was pressure, X2, the temperature and X3, the extraction time.
446 S. Zhao, D. Zhang / Separation and Purification Technology 133 (2014) 443–451

Table 2
Analysis of variance of the regression parameters.

Source Degree of freedom Regression coefficient F-values Probability (p) Significances

Model 14 22.91 <0.0001 Significant
Lack of fit 10 0.0507 Not significant
R2 = 0.9537
X1 pressure 1 0.78 127.30 <0.0001
X2 temperature 1 0.18 7.06 0.0240
X3 time 1 0.35 25.27 0.0005
X1X2 1 0.32 10.38 0.0091
X1X3 1 0.01 0.02 0.8813
X2X3 1 0.11 1.32 0.2767
X21 1 0.32 33.26 0.0002
X22 1 0.05 0.84 0.3807
X23 1 0.13 5.50 0.0410

The significance of each coefficient was also determined by the
F-values and p-values as listed in Table 2. A value of p < 0.05 indi-
cates that the equation term was significant. In this case, X1, X2, X3,
X1X2, X21, and X23 were all significant equation terms. Besides, the
larger magnitude of F-values and the smaller p-values mean a
higher significance of the corresponding coefficient. The influence
on the oil yield decreased in the order of pressure > extraction
time > temperature based on the F-values and p-values shown in
Table 2. Eliminating the terms that are not significant in the second
order polynomial equation, the prediction equation may be rear-
ranged to give the following equation (in terms of coded levels):
Y ¼ 4:12 þ 0:78X 1 þ 0:18X 2 þ 0:35X 3 þ 0:32X 1 X 2  0:32X 21
 0:13X 23
The best way to visualize the influence of the independent variables
on the oil yield as the dependent variable is to generate 3-dimen-
sional response surface plots by varying two variables while holding
the other two constants, as shown in Figs. 1 and 2.
Fig. 1 describes the interactive effect of pressure and tempera-
ture on the oil yield. The pressure and temperature are the main
parameters that influence the extraction efficiency. It was observed
that at a given temperature, the oil yield significantly increased
with increasing pressure, especially at low pressure levels. This is
because raising the extraction pressure at a constant temperature
led to a higher fluid density thus increased the solubility of the
Eucalyptus oil. It is also noted that the increase in oil yield with
increasing pressure became more apparent as the temperature
increased. For example, as pressure increased from 20 MPa to
40 MPa, the oil yield increased from 2.99% to 3.39% at 50 °C and
from 2.51% to 4.66% at 70 °C. Such increments in yields are due
to the interactions between the pressure and temperature. How-
ever, there was a negative quadratic effect at high pressures as
shown in Table 2. This is probably a reflection of the increased
repulsive solute–solvent interactions resulted from the highly
compressed CO2 at high pressures [18]. Furthermore, although at
a constant temperature, the increase in pressure would cause an
increase in fluid density, the fluid diffusion coefficient would
decrease. These counteracting effects result in higher pressures
having little effect on the oil yield [19,20]. Thus, a high pressure
is not always recommended for SFE.
It was very interesting to note that the effect of temperature on
ð4Þ
the oil yield showed different trends at different pressures. On one
hand, increasing temperature reduces the density of CO2 and thus
its solvation power at a constant pressure. On the other hand, rais-
ing temperature increases the solute vapour pressure, resulting in
enhanced SC-CO2 solubility. Therefore, the solubility of SC-CO2
may increase, remain constant, or decrease with rising tempera-
ture at a constant pressure, depending on whether solute vapour
pressure or solvent density dominates at the particular pressure
[19]. From Fig. 1, it can be seen that at low pressures, increasing
temperature resulted in a decrease in the oil yield. However at
higher pressures the oil yield increased with temperature. The oil
yield decreased from 3.50% to 3.06% when temperature was
increased from 50 to 70 °C at 20 MPa. However at 40 MPa the
same temperature increment corresponded to a change in the oil

Fig. 1. Response surface for the oil yield as related to temperature and pressure at Fig. 2. Response surface for the extracts as related to pressure and time at
fixed 90 min extraction. temperature of 60 °C.
 S. Zhao, D. Zhang / Separation and Purification Technology 133 (2014) 443–451 447

yield from 4.45% to 4.78%. This is because the solvent density

reduction at lower pressures was more influential than that in
higher pressures. For example, the density of supercritical CO2
reduced 0.13 g/ml from 50 to 70 °C at 20 MPa, while the reduction
was only 0.07 g/ml at 40 MPa. A similar trend was also reported by
Zhao and Zhang [12] in an investigate into the supercritical fluid
extraction of M. oleifera seeds oil using the response surface meth-
ods. It was found that there was a crossover pressure in the extrac-
tion process, when the extraction performed at pressures lower
than the crossover pressure, the yield of M. oleifera seeds oil
decreased with increasing temperature. However, when the pres-
sure was above the crossover pressure, the yield increased with
increasing temperature [12].
Fig. 2 presents the response surface plots, showing the effect of
pressure and extraction time on the oil yield at a fixed temperature
of 60 °C. How pressure affected the oil yield has been described in
Fig. 1. From Fig. 2, the extraction time exhibited an important
effect on the extraction efficiency, clearly due to the time required
for supercritical CO2 to penetrate into the substrate matrix, dis-
Fig. 3. Effect of extraction time on the yield of hydro-distillation of Eucalyptus leave
solve the oil and subsequently diffuse out from the material. A
oil.
longer extraction time had a positive linear effect on the oil yield.
However, for excessively long extraction time, the negative qua-
dratic effect also became significant. As shown in Fig. 2, a further of the extracts obtained by the two solvents were quite different;
increase in the extraction time from 120 min to 150 min resulted the hexane extracts contained a lot of alkanes and other hydrocar-
in little change in the oil yield. bons that were not detected in the ethanol extracts, such as octane,
Although there were some other factors which were not studied undecane, dodecane, heptadecane, nonadecane, eicosane, doco-
here such as the sample particle size and the solvent flow rate, the sane, toluene and benzene,1,3-dimethyl-. On the other hand, there
authors have reported them in other Refs. [12,13], In these previ- were a number of polar components that were identified in
ous studies, it was found that the oil yield increased with decreas- the ethanol extracts were not present in the hexane extracts,
ing particle size if the intra-particle diffusion had taken control of such as Myrtenol, .(-)-Carvone, 1,2,3-Benzenetriol, Cadinene, c-
the oil transfer during the extraction process. Although the oil yield Muurolene, p-Menthane-1,2,3-triol, d-Cadinene, Tetradecanoic
increased with increasing the solvent flow rate at the same acid, n-Hexadecanoic acid, 9,12,15-Octadecatrienoic acid, (Z,Z,Z)-
extraction time, the solubility of oil was not a function of CO2 flow and c-Sitosterol. It is well known that hexane is a non-polar
rate [13]. solvent while ethanol is a strong polar solvent. Based on the basic
principle of ‘‘like dissolves like’’ [21], the alkane compounds are
3.2. Hydro-distillation non-polar and therefore more easily extracted by hexane. For the
polar compounds such as Myrtenol, 1,2,3-Benzenetriol et al., as
Fig. 3 shows Eucalyptus oil yield as a function of extraction time they contain the O–H ends, they readily interact with the hydrogen
in hydro-distillation. As expected, the oil yield increased with bond of ethanol and thus make them more easily separated by
increasing time, from 3.1% at 1 h to 3.8% at 5 h of extraction, indi- ethanol. This observation in the composition details not only
cating a rapid extraction rate in the first 1 h. This is probably explains the huge difference in the ethanol and hexane Soxhlet
because the Eucalyptus leaves grinding process liberated the essen- extraction yields, it also implies that essential oils often extracted
tial oil onto the surface of the plant particles [14], allowing more using such methods are not always the same.
ready extraction by the water vapour. Therefore, most of oil was
extracted during the first hour [14].

3.3. The effect of operating parameters in Soxhlet extraction

To compare the SFE with other common extraction methods,

the effect of solvent type and extraction time on the oil yield
obtained by Soxhlet extraction was also evaluated and the results
are shown in Fig. 4. The experimental results were the average of
three measurements, with error bars showing the standard devia-
tion of these measurements. Ethanol and hexane were chosen in
this study as the extraction solvent. Fig. 4 shows that the oil yield
rapidly increased with increasing extraction time till ca. 2 h, but
levelled off with further increasing from 2 h to 8 h. However, the
solvent type had a more significant effect on the oil yield, with eth-
anol showing much higher oil yield than hexane over the same
extraction time. For example, the oil yield extracted by ethanol
at 8 h was 36.33% which was nearly four times that by hexane at
7.90%. In order to explain the difference, the total ion compound
chromatograms of the Eucalyptus leaves extracts obtained by the
two solvents were acquired using a GC–MS. The results are shown
in Fig. 5 and the detailed compounds information is listed in the Fig. 4. Effect of solvent type and extraction time on the yield of Soxhlet extraction
Supplementary information. It is obvious that the compositions of Eucalyptus leaves oil.
448 S. Zhao, D. Zhang / Separation and Purification Technology 133 (2014) 443–451
Fig. 5. Total ion compound (TIC) chromatograms of the Eucalyptus leaves extracts
obtained by Soxhlet extraction with ethanol and hexane.
3.4. Comparison of SFE, Soxhlet extraction and hydro-distillation
methods
The oil yields of Eucalyptus leaves obtained by SFE under all con-
ditions, hydro-distillation and Soxhlet extraction with ethanol and
hexane are summarised in Fig. 6. It can be seen that Soxhlet extrac-
tion with ethanol brought out the highest oil yield of 36.33%, while
hydro-distillation had the lowest yield of 3.77% and the SFE
showed an intermediate yield. The SFE yield was up to 4.78%, com-
parable in magnitude with the hexane Soxhlet results. Similar
results have been reported by other authors [22,23]. Glisic et al.
[22] investigated the SC-CO2 extraction of Dalmatian sage and
compared that with Soxhlet extraction using ethanol–water
(70:30) mixture and hydro-distillation. Their results revealed that
the Soxhlet extraction achieved the highest yield of 26.5 wt%, while
SC-CO2 extraction at 30 MPa resulted in a yield of 4.82 wt% and
hydro-distillation only gave a yield of 0.5% (v/w) of essential oil.
They explained that the SFE allowed a wide range of chemicals
to be isolated, while hydro-distillation was limited to volatile
compounds and Soxhlet extraction was limited to high molecular
compounds. Vági et al. [24] have also compared the extracts from
marjoram using SC-CO2 (50 °C and 45 MPa), ethanol Soxhlet
extraction and hydro-distillation and the oil yields were 3.8 g/
100 g, 9.1 g/100 g, and 0.8 ml/100 g, respectively. However, the
SFE comprised 21% essential oil, while the alcoholic extract
contained only 9% volatile oil substance. Zhao and Zhang [25] also
compared the supercritical fluid extraction of M. oleifera leaves
oil with the Soxhlet hexane extraction. The yield of the SFE process
(the highest yield was 6.3%) was lower than that of the Soxhlet
hexane extraction (9.3%).
Fig. 6. Extraction yields of Eucalyptus leaves oil obtained by different methods:
hydro-distillation (HD), Soxhlet extraction with hexane at different times (hexane),
Soxhlet extraction with ethanol at different times (ethanol) and supercritical fluid
extraction (SFE) under different extraction conditions.
In order to explain the differences in the oil yields from the
three extraction methods, the oil samples were analysed using
GC–MS for their compositions. The chemical profiles of ethanol
and hexane Soxhlet extracts have been discussed in Section 3.3
and are shown in Fig. 5. The GC–MS profiles of SFE (under condi-
tions of 40 MPa, 70 °C and time 1 h) and hydro-distillation oils
are shown Fig. 7. The major compounds identified are listed in
the Supplementary information in detail. The fractions of all
compounds were calculated using the area normalization method
[26]. Clearly, the chemical compositions of the oil extracted by
the three methods differed both qualitatively and quantitatively.
In order to appreciate the chemical compositions of the Eucalyptus
leaves oil, the identified compounds are grouped in Fig. 8 into five
compound families: monoterpenes, oxygenated monoterpenes,
sesquiterpenes, oxygenated sesquiterpenes and other compounds.
Fig. 8 shows that the common compounds present in the Euca-
lyptus oils obtained by all three methods were oxygenated mono-
terpenes, sesquiterpenes and oxygenated sesquiterpenes. The
main oxygenated monoterpenes, 1,8-cineole, the key and most
beneficial ingredient of essential oil, was consistently the most
abundant compound in every sample. The proportions of this
constituent in the extracted oils were 46.19% for SFE, 70.03% for
hydro-distillation sample, 29.85% for ethanol Soxhlet, and 56.4%
for hexane Soxhlet. Other major oxygenated monoterpenes
present in all the oil samples, but at much lower concentrations,
were L-pinocarveol, Pinocarvone, 4-Terpineol, p-Cymen-8-ol, a-
Terpinol, Verbenone, cis-Carveol, cuminal, pipertitone, L-perillal-
dehyde and Thymol. Oxygenated sesquiterpenes was the second
most significant group of compounds in all the oil samples, includ-
ing Epiglobulol, Palustrol, (-)-Spathulenol, Caryophyllene oxide,
Globulol, Germacrene D-4-ol, Ledol, Cubenol, tau.-Muutolol, d-
Cadinol, Isoaromadendrene epoxide and Longiverbenone. The third
major group was the sesquiterpens substances such as Copaene, b-
Cubebene, (+)-Aromadendrene and Cadalene. Although some
monoterpene hydrocarbons like o-Cymol were also detected, but
only in trace amounts found in the SFE and Soxhlet extracts and
in small amounts in the hydro-distillation oil.
The GC–MS results show that the number of compounds found
in the hydro-distillated oil was significantly less than those in the
SFE and Soxhlet extracts. This is because the Eucalyptus oil
extracted by hydro-distillation contained mostly volatile com-
pounds, with the majority appearing in the first 66 min on the
GC separation chromatograph as shown in Fig. 7. In contrast, The
 S. Zhao, D. Zhang / Separation and Purification Technology 133 (2014) 443–451
Fig. 7. Total ion compound (TIC) chromatograms of the Eucalyptus leaves oil
obtained by supercritical fluid extraction and hydro-distillation.
Fig. 8. Composition profile of Eucalyptus leaves oil extracted by SFE (under
conditions of 40 MPa, 70 °C and time 1 h), hydro-distillation, and Soxhlet extraction.
SFE and Soxhlet oils contained not only the volatile components,
but also heavier compounds appearing after 66 min on the GC
separation chromatograph (Fig. 7).
Similar results have been reported in the literature [7,8]. da
Cruz Francisco et al. [7] compared the Eucalyptus camaldulensis
Dehn. oils obtained by hydro-distillation and supercritical carbon
dioxide extraction. The oil obtained by hydro-distillation showed
a high concentration of 1,8-cineole (43.4%), a-pinene, b-pinene,
p-cymene, terpinene-4-ol and glubulol. However, the oils obtained
by supercritical carbon dioxide extraction had lower amounts of
449
Hydro-distillation process:
1.the desorption of the substrate
2. the internal diffusion of the substrate
3 3. Azeotropic (water-essential oil)
distillation
Soxhlet extraction process:
1.the diffusion of the solvent
2.the dissolution of extracted substance
4 3. the internal diffusion of the substrate
and the solvent
4. the external diffusion of the substrate
3
and the solvent
2
4
Supercritical extraction process:
1. the diffusion of the supercritical fluid
3
2. the dissolution of extracted substance
2 3. the diffusion of the substrate and the
supercritical fluid
1 4. flow away of the supercritical fluid and
the substrate mixture
Fig. 9. Schematic diagrams of the hydro-distillation process, Soxhlet extraction
process and supercritical fluid extraction process.
1,8-cineloe, a-pinene, b-pinene, p-cymene, terpinene-4-ol, but
higher amounts of high molecular compounds such as allo-aroma-
dendrene and globulol. This is similar to the current finding that
the oil extracted by SFE contented less 1,8-cineole compared to
that extracted by hydro-distillation, and that supercritical CO2
extracted high molecular weight compounds as well. It was
explained [7] that high-molecular weight compounds including
esters, fatty acids, and waxy-like compounds are more likely to
be extracted by CO2 rather than by hydro-distillation which more
easily extract the volatile compounds [7]. However, the opposite
observations were reported in Della Porta et al. [8] who found that
the 1,8-cineole content was higher in supercritical carbon dioxide
extraction (62.6%) than in hydro-distillation(48.2%). This was
because they performed two fractions during the extraction
process. The volatile oil was recovered in the second separator
and therefore had higher cineole content, whereas waxes was
recovered in the first separator. They suggested that the optimum
fraction would be achieved by operating at 90 bar and 10 °C in
the first separator and 15 bar and 15 °C in the second.
The reason for the observed differences in composition could be
due to their different isolation mechanisms as shown in Fig. 9. In
the hydro-distillation process, water does not dissolve organic
components. The substrate sample is immersed in water and then
heated to ca 100 °C. The essential oil components desorb from the
substrate matrix and then diffuse into the liquid phase. These com-
ponents form an azeotropic mixture with water, then the volatile
constitutes gradually evaporate from the extraction liquid into
the steam phase, which then condense in the condenser. The oil
separates from water to form an oil phase. The water is returned
to the flask to hydro-distillate the sample again. In this process,
only volatile compounds can be evaporated and carried away by
450 S. Zhao, D. Zhang / Separation and Purification Technology 133 (2014) 443–451

steam. The heavier compounds with higher boiling points are hard
to evaporate with water and thus are not detected in the hydro-
distillation oil as observed in this study. In conventional Soxhlet
extraction, an organic solvent of hexane and ethanol is used. The
solvent type and properties have a strong influence on the compo-
sitions of extracts from the substrate. Hexane Soxhlet achieved
higher oil yields than SFE and hydro-distillation in this work. On
the other hand, ethanol Soxhlet achieved the highest oil yield, by
an order of magnitude higher than those of the other methods
studied in this work. As aforementioned, this is because ethanol,
owing to its polar nature, can extract not only the polar compounds
that are not dissolvable in hexane, but also non-polar components.
Besides, ethanol Soxhlet also extracts very heavy components such
as waxes, chlorophyll and others that were not detected in the
GC–MS analysis [27]. Furthermore, the contact between the sol-
vent and the substrate in Soxhlet extraction is different from those
in hydro-distillation and supercritical fluid extraction. In Soxhlet
extraction, the solvent is heated to reflux, the vapour passes
Fig. 10. Effect of ethanol’s concentration on the yield of Eucalyptus leaves oil.
through a bypass arm to reach the condenser. Then the solvent
drips down into the thimble and diffuses into the substrate matrix. findings. Michielin et al. [28] studied the extraction of Cordia
The extractable substances then dissolve into the solvent. The mix- verbenacea using pure CO2 and CO2 with modifiers of ethanol and
ture of the solvent and extracted substances then diffuses from the ethyl acetate, and the yield increased up to 54% using 5% ethanol
substrate matrix back to the bulk solvent. Once the mixture level in compared to pure CO2. They concluded that these increases
the extractor reaches the top of the siphon arm, the solvent and the were related to the extraction of the more polar solutes from the
extract are siphoned back into the lower flask. Thus, Soxhlet substrates.
extraction is a batch extraction process. SC-CO2 was used in the The compositions of the Eucalyptus leaves oils extracted using
supercritical fluid extraction in the present study. The oil yields pure supercritical CO2 and that with ethanol as a modifier were
in SFE and hexane Soxhlet were of the same magnitude. This is analysed using GC–MS and the detailed compounds information
so because both SC-CO2 and hexane are non-polar solvents and obtained is listed in the Supplementary Information. As 1,8-cineole
they only preferably solubilize non-polar components. However, is the most important components in the Eucalyptus leaves oil
the molecular weight of hexane is greater than that of CO2, which determine the oil quality and market value, the effect of
enabling it to dissolve some long chain alkanes or heavy molecular ethanol concentration on the content of 1,8-cineole was also stud-
compounds more easily than SC-CO2. Therefore the compositions ied and the results are shown in Fig. 11. It can be seen that the eth-
of the oil extracted were quite different when hexane Soxhlet anol concentration had a significant effect on the 1,8-cineole
extraction and supercritical fluid extraction were employed, content in the Eucalyptus leaves oil extracted at 10 MPa. The 1,8-
respectively. In the supercritical fluid extraction process, once cineole content decreased significantly from 77.0% to 46.1% with
SC-CO2 diffuses into the substrate matrix and solubilises the increasing ethanol concentration from 5 w% to 15 w%. This is
extractable compounds, the mixture of the compounds and SC-CO2 because the solubility of 1,8-cineole in the ethanol modified sol-
diffuses out of the substrate matrix and flows away with the bulk vent decreased with increasing ethanol concentration at 10 MPa.
fluid phase. Therefore supercritical fluid extraction is a continuous The effects of modifiers including ethanol, water, acetone and hex-
process, involving very different transport phenomena from hydro- ane on the solubilities of a-pinene and 1,8-cineole in supercritical
distillation and Soxhlet extraction processes, as shown in Fig. 9. CO2 have been investigated by Shimoyama et al. [29] at 50 °C and
Comparing these three techniques, it can be seen that although 8.0 MPa. They found that the solubility of 1,8-cineole decreased
the oil extracted by hydro-distillation had the highest content of from 0.00853 mol/L to 0.00736 mol/L with ethanol concentration
1,8-cineole, this method produced the lowest yield and it took long increased from 0.00221 mol/L to 0.00562 mol/L at this low
extraction times to achieve its maximum yield. The oil yield pressure. However, on the other hand, there was no significant
obtained by ethanol Soxhlet was the highest but had a very low
1,8-cineole content. The SFE is therefore considered the best alter-
native method for Eucalyptus oil extraction.

3.5. The effect of modifier on SFE

As the Soxhlet extraction using ethanol produced the highest

Eucalyptus leaves oil yield, the introduction of ethanol in the
supercritical CO2 extraction of Eucalyptus leaves oil was investi-
gated. The oil yield at ethanol concentrations ranging from 5 wt%
to 15 w% was shown in Fig. 10. Similar to the discussion before,
the experiments conducted at higher pressure (30 MPa) resulted
in a higher yield than at low pressure (10 MPa), regardless of
ethanol concentration. Besides, it can be seen that the yield of
Eucalyptus leaves oil increased with increasing ethanol concentra-
tion at both high and low pressures. This is because the polarity of
the mixed solvent increased with increasing ethanol concentration,
as most of the compounds in Eucalyptus leaves are polar com-
pounds and thus they are more likely to be dissolved in highly
polar solvents. Some other researchers have also observed similar Fig. 11. Effect of ethanol’s concentration on the content of 1,8-cineole.
 S. Zhao, D. Zhang / Separation and Purification Technology 133 (2014) 443–451
difference in the of 1,8-cineole content, almost invariant at ca 49%,
at 30 MPa when the ethanol concentration was varied. No informa-
tion on the solubility of 1,8-cineole in ethanol modified supercrit-
ical CO2 under high pressures has been reported in the literature. It
is thus surmised that the unchanged content could be attributed to
the greater solvating power of supercritical CO2 at the high pres-
sures. At 30 MPa, the density of supercritical CO2 is very high
and its effect exceeded that of ethanol, hence the 1,8-cineole con-
tent did not change significantly with the introduction of ethanol.
4. Conclusions
Supercritical fluid extraction was successfully applied to extract
oil from Eucalyptus leaves using CO2 and the effect of operating
pressure, temperature and extraction time on the oil yield were
studied using a central composition design methodology. Oil yield
increased with increasing pressure and extraction time. However,
temperature had a moderate quadratic effect and interacted with
the pressure effect. A simplified regression equation was devel-
oped to predict the Eucalyptus leaves oil yield and was shown to
be sufficient to describe the extraction process. The oil yield and
chemical compositions of the oils obtained by SFE, hydro-
distillation and Soxhlet extraction were also compared. In the
Soxhlet extraction, the solvent type had a more significant effect
on the oil yield than the extraction time, while in hydro-
distillation, the oil yield increased with extraction time. The
hydro-distillation gave the lowest yield of 3.77% while the ethanol
Soxhlet extraction provided the highest yield of Eucalyptus leaves
oil at about 36.33%. It was also noted that due to the different
mechanisms of the three methods, hydro-distillation had only
extracted the volatile compounds, while Soxhlet extraction and
SFE had extracted a wide range of components including not only
the volatile but also the high molecular compounds. The content of
1,8-cineole in the extracted oils were 46.19% for SFE, 70.03% for
hydro-distillation sample and 29.85% for ethanol Soxhlet sample.
The addition of ethanol as a modifier in supercritical CO2 improved
the efficiency of supercritical fluid extraction. The yield of Eucalyp-
tus leaves oil increased with increasing ethanol concentration. The
compositions of the oil also differed quantitatively at different
ethanol concentrations at low extraction pressures. There was a
significant decrease in the 1,8-cineole content with increasing
ethanol concentration from 5 w% to 15 w% at 10 MPa.
Acknowledgments
The authors gratefully acknowledge the financial and other sup-
port provided by the Australian Research Council under the ARC
Linkage Projects Scheme (LP100200135), BHP Billiton Iron Ore
Pty Ltd and ANSAC Pty Ltd.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.seppur.2014.07.
018.
References
[1] G.D.K. Babu, B. Singh, Simulation of Eucalyptus cinerea oil distillation: a study
on optimization of 1,8-cineole production, Biochem. Eng. J. 44 (2009) 226–231.
[2] D.R. Batish, H.P. Singh, R.K. Kohli, S. Kaur, Eucalyptus essential oil as a natural
pesticide, For. Ecol. Manage. 256 (2008) 2166–2174.
451
[3] M. Gilles, J. Zhao, M. An, S. Agboola, Chemical composition and antimicrobial
properties of essential oils of three Australian Eucalyptus species, Food Chem.
119 (2010) 731–737.
[4] J.d.C. Francisco, B. Sivik, Solubility of three monoterpenes, their mixtures and
eucalyptus leaf oils in dense carbon dioxide, J. Supercrit. Fluids 23 (2002) 11–
19.
[5] D.G. Ammon, A.F.M. Barton, D.A. Clarke, J. Tjandra, Rapid and accurate
determination of terpenes in the leaves of Eucalyptus species, Analyst 110
(1985) 921–924.
[6] E. Reverchon, I. De Marco, Supercritical fluid extraction and fractionation of
natural matter, J. Supercrit. Fluids 38 (2006) 146–166.
[7] J. da Cruz Francisco, E.P. Järvenpää, R. Huopalahti, B. Sivik, Comparison of
Eucalyptus camaldulensis Dehn. Oils from Mozambique as obtained by
hydrodistillation and supercritical carbon dioxide extraction, J. Agric. Food
Chem. 49 (2001) 2339–2342.
[8] G. Della Porta, S. Porcedda, B. Marongiu, E. Reverchon, Isolation of eucalyptus
oil by supercritical fluid extraction, Flavour Frag. J. 14 (1999) 214–218.
[9] H.S. Yim, F.Y. Chye, S.M. Koo, P. Matanjun, S.E. How, C.W. Ho, Optimization of
extraction time and temperature for antioxidant activity of edible wild
mushroom, Pleurotus porrigens, Food Bioprod. Process. 90 (2012) 235–242.
[10] M.J. Ramalhosa, P. Paíga, S. Morais, A.M.M. Sousa, M.P. Gonçalves, C. Delerue-
Matos, M.B.P.P. Oliveira, Analysis of polycyclic aromatic hydrocarbons in fish:
optimization and validation of microwave-assisted extraction by response
surface methodology, Food Chemistry.
[11] F. Li, L. Yang, T. Zhao, J. Zhao, Y. Zou, Y. Zou, X. Wu, Optimization of enzymatic
pretreatment for n-hexane extraction of oil from Silybum marianum seeds
using response surface methodology, Food Bioprod. Process. 90 (2012) 87–94.
[12] S. Zhao, D. Zhang, A parametric study of supercritical carbon dioxide extraction
of oil from Moringa oleifera seeds using a response surface methodology, Sep.
Purif. Technol. 113 (2013) 9–17.
[13] S. Zhao, D. Zhang, An experimental investigation into the solubility of Moringa
oleifera oil in supercritical carbon dioxide, J. Food Eng. 138 (2014) 1–10.
[14] M. Gavahian, A. Farahnaky, K. Javidnia, M. Majzoobi, Comparison of ohmic-
assisted hydrodistillation with traditional hydrodistillation for the extraction
of essential oils from Thymus vulgaris L., Innovat. Food Sci. Emerg. Technol. 14
(2012) 85–91.
[15] S. Zhao, D. Zhang, Supercritical fliud extraction and characterisation of
essential oil from Moringa olefera leaves. In: Chemeca 2011, Sydney, 2011.
[16] S. Stigler, Fisher and the 5% level, Chance 21 (2008).
[17] H.-Y. Cui, X.-Y. Jia, X. Zhang, J. Zhang, Z.-Q. Zhang, Optimization of high-speed
counter-current chromatography for separation of polyphenols from the
extract of hawthorn (Crataegus laevigata) with response surface methodology,
Sep. Purif. Technol. 77 (2011) 269–274.
[18] S. Liu, F. Yang, C. Zhang, H. Ji, P. Hong, C. Deng, Optimization of process
parameters for supercritical carbon dioxide extraction of Passiflora seed oil by
response surface methodology, J. Supercrit. Fluids 48 (2009) 9–14.
[19] H. Wang, Y. Liu, S. Wei, Z. Yan, Application of response surface methodology to
optimise supercritical carbon dioxide extraction of essential oil from Cyperus
rotundus Linn, Food Chem. 132 (2012) 582–587.
_ S
[20] I.S. ß anal, E. Bayraktar, Ü. Mehmetoğlu, A. Çalımlı, Determination of optimum
conditions for SC-(CO2 + ethanol) extraction of b-carotene from apricot
pomace using response surface methodology, J. Supercrit. Fluids 34 (2005)
331–338.
[21] M.K.L. Bicking, Extraction|Analytical Extractions, in: D.W. Ian (Ed.) Editor-
in-Chief, Encyclopedia of Separation Science, Academic Press, Oxford, 2000,
pp. 1371–1382.
[22] S. Glisic, J. Ivanovic, M. Ristic, D. Skala, Extraction of sage (Salvia officinalis L.)
by supercritical CO2: kinetic data, chemical composition and selectivity of
diterpenes, J. Supercrit. Fluids 52 (2010) 62–70.
[23] W. Guan, S. Li, R. Yan, S. Tang, C. Quan, Comparison of essential oils of clove
buds extracted with supercritical carbon dioxide and other three traditional
extraction methods, Food Chem. 101 (2007) 1558–1564.
[24] E. Vági, B. Simándi, Á. Suhajda, É. Héthelyi, Essential oil composition and
antimicrobial activity of Origanum majorana L. extracts obtained with ethyl
alcohol and supercritical carbon dioxide, Food Res. Int. 38 (2005) 51–57.
[25] S. Zhao, D. Zhang, Supercritical fluid extraction and characterisation of Moringa
oleifera leaves oil, Sep. Purif. Technol. 118 (2013) 497–502.
[26] S. Zhao, D. Zhang, A parametric study of supercritical carbon dioxide extraction
of oil from Moringa oleifera seeds using a response surface methodology, Sep.
Purif. Technol.
[27] L. Danielski, L.M.A.S. Campos, L.F.V. Bresciani, H. Hense, R.A. Yunes, S.R.S.
Ferreira, in: Marigold (Calendula officinalis L.) oleoresin: solubility in SC-CO2
and composition profile, Chem. Eng. Process. 46 (2007) 99–106.
[28] E.M.Z. Michielin, A.A. Salvador, C.A.S. Riehl, A. Smânia Jr., E.F.A. Smânia, S.R.S.
Ferreira, Chemical composition and antibacterial activity of Cordia verbenacea
extracts obtained by different methods, Bioresour. Technol. 100 (2009) 6615–
6623.
[29] Y. Shimoyama, H. Tokumoto, T. Matsuno, Y. Iwai, Analysis of cosolvent effect
on supercritical carbon dioxide extraction for a-pinene and 1,8-cineole, Chem.
Eng. Res. Des. 88 (2010) 1563–1568.