Natural Science, 2009, 1, 23-29 NS
http://dx.doi.org/10.4236/ns.2009.11005
Optimization of solvent extraction conditions for total
carotenoids in rapeseed using response surface
methodology
Ling Wang1, Yun Liu1,*
1College of Life Science and Technology, Huazhong University of Science and Technology, Wuhan 430074, P. R. China.
Received 10 May 2009; revised 16 May 2009; accepted 19 May 2009.
ABSTRACT
The optimum total carotenoids (TC) extraction
from rapeseed with solvent extraction method
by UV-visible spectrophotometer determination
was investigated by using response surface
methodology (RSM). Extraction duration, re-
peated extraction cycles, solvent-solid ratio and
extraction temperature were assumed to be the
most important factors affecting solvent extrac-
tion for the determination of TC. Optimum sol-
vent extraction conditions for maximizing the
determination of TC were: extraction duration
7.3h, repeated extraction three times, ratio of
solvent-solid (v/w, mL/mg) 29:1, extraction
temperature 42°C. Under the optimal conditions,
the yield of TC was up to 4.79 mg /100g. The
model had a satisfactory coefficient of R2 (=
0.912) and verified experimentally. The results
showed that the conditions were mild and use-
ful for maximizing a quantitative spectropho-
tometer determination of TC in rapeseed.
Keywords: Carotenoids; Optimization; Solvent Ex-
traction; Response Surface Methodology; Rape-
seed
1. INTRODUCTION
Carotenoids are a group of phytochemical bioactive
compounds that are responsible for different colors of
various plants and microorganisms but not animals [1].
It has been found that carotenoids can play an important
role in the prevention of various types of cancer as well
as other important ‘‘lifestyle- related’’ diseases, such as
cardiovascular disease and age-related macular degen-
eration due to their antioxidant activity [2,3,4]. In addi-
tion to being potent antioxidants some carotenoids also
contribute to provitamin A function [5]. Although the
chemistry properties of carotenoids have been exten-
sively studied their bioavailability, metabolism and bio-
logical functions are only investigated recently [6,7]. In
recent years the antioxidant properties of carotenoids
have become the major focuses for researches, particu-
larly focused on the role of lycopene in human health
[8,9]. About 90% of the carotenoids in the diet and hu-
man body are represented by β-carotene, α-carotene,
lycopene, lutein and cryptoxanthin [10].
As the increasing of health-conscious and the demand
for carotenoids, researchers shifted their attentions from
chemical synthesis to natural products isolated from
plants and microorganisms biological sources [11,12].
Rapeseed containing around 40% oil is one of the most
important vegetable oil materials in the world. The total
production of rapeseed plant all over the world was 46.2
Mt in 2005 [13]. Rapeseed contains rich carotenoids,
such as β-carotein, α-carotein and lutein, which can con-
tribute to prolong the rapeseed oil shelf life and increase
oil nutrition [14,15]. Many effective methods for caro-
tenoids extraction from biological sources have been
intensively employed, such as solvent extraction, solid
phase extraction (SPE) and supercritical fluid extraction
(SFE) [12,16,17]. Above these methods, solvent extrac-
tion method is universally application for extraction total
carotenoids (TC) because solvent extraction method is
relative simple and low cost [18,19,20,21]. However, the
extraction of TC from rapeseed was rarely reported so
far.
The most commonly used techniques for TC detection
are UV-visible spectrophotometry, mass spectrometry
and hydrogen or carbon nuclear magnetic resonance
spectroscopy (1H-NMR, 13C-NMR), coupled or not with
chromatographic techniques [22,23]. Regardless of the
technique utilized, carotenoid extraction is highly influ-
enced by procedural variables such as type of sample,
type and ratio of solvents, extraction duration, repeated
extraction times, storage conditions, etc. Therefore, the
objective of this work was to establish a solvent extrac-
tion method for a quantitative spectrophotometer deter-
mination of TC. Rapeseed was selected as representative
due to its abundance all over the world. Response sur-

24 L. Wang et al. / Natural Science 1 (2009) 23-29
face methodology (RSM) was employed to optimize the
extraction conditions, which could maximize the deter-
mination of TC in rapeseed.
2. MATERIALS AND METHODS
2.1 Materials
Rapeseed (Brassica napus L.) was provided by Rapeseed
Engineering Center of Huazhong University of Agricul-
ture. The raw material consisted of moisture 2.9 ± 0.05%,
kernel 73.4 ± 0.26%, and seed capsule 26.6 ± 0.02%.
Hexamethylene, petroleum ether, chloroform, acetone
and methanol were of analytic grade.
2.2 Experimental Design
Six extraction solvents of hexamethylene, petroleum ether,
chloroform, acetone, methanol and mixture of petroleum
ether and acetone were tested to select the most optimum
extraction solution for TC extraction. 3-5 g rapeseed was
grinded into powder in a glass mortar and screened with
100 mesh sieve. Portions of ground rapeseed powder of
50mg (dry weight) were transferred to 40mL beakers
added with 5mL extraction solvent and wrapped with
aluminum foil. The samples were constantly agitated (Si-
satom magnetic agitator) according to the extraction dura-
tion, protected from light at room temperature (23°C). In
the preliminary study, variables affecting TC extraction
were solid-solvent ratio, extraction duration, extraction
repeated cycles and extraction temperature. Response
surface methodology (RSM) was used to optimize the
above parameters. A four-factor- five-level centre com-
posite design was adopted to optimize the extraction con-
ditions for analysis of the TC in rapeseed. The quadratic
response surface model fitted Eq.1:
0
bY +

k
i
Xibi
1
+

k
i
i
Xbii
1
2+ji
k
i
k
ij
ij XXb

1
+e
(1)
where Y standed for the total carotenoids yield, b0 de-
noted the model intercept, i and j were the linear and
quadratic coefficients, respectively, bi, bii and bij were
the regression coefficient, k was the number of factors
studied and optimized in the experiment and e was the
random error. Statistical Analysis System (SAS Institute
Inc, Cary, NC, USA) was used to fit the second order
polynomial equation to the experimental data.
The goodness of fit of the model was evaluated by the
coefficient of determination (R2) and the analysis of
variance (ANOVA). Quadratic polynomial equations
were attained by holding one of the independent vari-
ances at a constant value and changing the level of the
other variables.
2.3 Total Carotenoids Quantification
After extraction, the samples were filtered on filter paper,
their volume was made up to 3mL, and they were stored
in an amber flask (~10mL) filled with N2. To determine
the amount of total carotenoids extracted, UV-visible
spectrophotometer (UNICO UV-2802, USA) was used
for a spectral window between 380 and 750 nm, in trip-
licate. The absorbance value of carotenoids extract was
monitored at 445 nm. The total carotenoids (TC) yield
(mg/100g) was calculated according to the following Eq.2:
wA
mLyA
gmgTC
cm 

1000
10)(
)100/( %
1
6
(2)
where A was the absorbance value of extract at 445 nm,
y was the volume of extract, %
1cm
A was the extinction
coefficient of carotenoids, and w was the weight of
rapeseed powder (g).
2.4 Verification of Model
Optimizations of extraction conditions, including reac-
tion temperature, solid-solvent ration, extraction dura-
tion, and extraction repeated cycles for maximizing a
quantitative UV-visible spectrophotometer determination
of TC in the rapeseed were calculated by using the pre-
dictive equation from RSM. The actual determination of
TC was carried out by UV-visible spectrophotometer
after extraction at the optimum conditions, and the result
was compared to the predicted value.
3. RESULTS AND DISCUSSIONS
3.1. The Effect of Extraction Solvents
on TC Yield
Six different extraction solvents, hexamethylene, petro-
leum ether, chloroform, acetone, methanol and petro-
leum ether/acetone mixture, were employed to choose
the most suitable solvent for TC extraction from rape-
seed. The experimental results were shown in Table 1.
It was indicated from Table 1 that the mixture solvent
of petroleum ether and acetone (1:1, v/v) was the most
effective solvent for TC extraction from rapeseed. As
known, rapeseed contains polar carotenoid such as lutein,
and nonpolar carotenoids such as β-carotene and carote-
noids ester, the former is easily dissolved in polar sol-
vent (e.g. acetone) while the latter is easily dissolved in
nonpolar solvent (e.g. petroleum ether). Therefore, the
Table 1. Effects of extraction solvent on TC yield.
Extraction solvents TC yield (mg/100g)
Hexamethylene 0.617±0.065
Petroleum ether 0.972±0.038
Chloroform 2.110±0.052
Acetone 2.438±0.063
Methanol 2.312±0.075
Petroleum ether/acetone mix-
ture(1:1,v/v) 3.890±0.093

L. Wang et al. / Natural Science 1 (2009) 23-29 25
mixture solvent of petroleum ether and acetone was most
suitable solvent for the TC extraction from rapeseed
among the tested solvents.
3.2 Optimization of Extraction Conditions
by RSM
In our preliminary study, variables affecting TC extrac-
tion were solid-solvent ratio, extraction duration, extrac-
tion repeated cycles and extraction temperature. The
optimum experiments were conducted by using a five-
level-four-factor central composite design with twelve
replicates at the central point. The coded and actual lev-
els of the three variables in Table 2 were selected to
maximize the UV-visible spectrophotometer determina-
tion of total carotenoids (TC).
Table 2. Coded and actual levels of four variables.
Factors Duration
(X1, h)
Repeated
cycle (X2)
Solid-solvent
ratio
(X3, g/mL)
Temperature
(X4, °C)
r=2 10 5th 1:50 60
1 8 4th 1:40 50
0 6 3rd 1:30 40
-1 4 2nd 1:20 30
-r=-22 1st 1:10 20
△x 2 1 10 10
Table 3 showed the treatments with coded levels and
their experimental results of TC in rapeseed.
The TC yield ranged from 2.762mg/100g to 4.863mg
/100g, and the run=20 and the run=27 had the minimum
Table 3. Coded level combinations for a four-variable central composite orthogonal and rotatable design (CCD).
Runs X1 X2 X3 X4 TC(Y1, mg/100g)
1 -1 -1 -1 -1 3.046 ±0.061
2 -1 -1 -1 1 3.726±0.053
3 -1 -1 1 -1 3.192±0.047
4 -1 -1 1 1 4.006 ±0.035
5 -1 1 -1 -1 3.999±0.018
6 -1 1 -1 1 3.110 ±0.009
7 -1 1 1 -1 3.665±0.016
8 -1 1 1 1 3.898 ±0.024
9 1 -1 -1 -1 4.226±0.034
10 1 -1 -1 1 4.140±0.018
11 1 -1 1 -1 3.123±0.017
12 1 -1 1 1 2.760±0.005
13 1 1 -1 -1 3.314±0.032
14 1 1 -1 1 4.325±0.029
15 1 1 1 -1 4.348±0.034
16 1 1 1 1 3.935±0.015
17 -2 0 0 0 4.437±0.012
18 2 0 0 0 4.088±0.011
19 0 -2 0 0 4.542±0.065
20 0 2 0 0 2.762±0.002
21 0 0 -2 0 3.664±0.041
22 0 0 2 0 4.135±0.036
23 0 0 0 -2 3.159±0.025
24 0 0 0 2 3.834±0.015
25 0 0 0 0 4.768±0.023
26 0 0 0 0 4.805±0.025
27 0 0 0 0 4.863±0.031
28 0 0 0 0 4.763±0.015
29 0 0 0 0 4.822±0.028
30 0 0 0 0 4.719±0.017
31 0 0 0 0 4.768±0.021
32 0 0 0 0 4.805±0.018
33 0 0 0 0 4.343±0.014
34 0 0 0 0 4.712±0.010
35 0 0 0 0 4.802±0.034
36 0 0 0 0 4.779±0.016
and maximum yield, respectively. Using the designed
experimental data (Table 3), the polynomial model de-
scribing the correlation between TC yield and the four
variables or conditions was obtained as follows:

26 L. Wang et al. / Natural Science 1 (2009) 23-29
TC(Y1, mg/100g)=4.77896+0.034687*X1-0.049457 *X2
-0.00073*X3+0 .097412* X4-0.15435 3*X1*X1+0.060615
*X1*X2-0.169996*X1*X3-0.043019*X1*X4-0.307039*X
2*X2+0.197118*X2*X3-0.068865*X2*X4-0.245239*X3*
X3-0.027968*X3*X4-0.345873*X4*X4
Table 4 showed the analysis of variance (F-test) for
this model, and the coefficient of determination (R2) was
shown as 91.16%. The regression analysis showed that
91.16% of the variations were explained by the model.
This indicated that the accuracy and general availability
of the polynomial model was good, analysis of the re-
sponse trends using the model was considered to be rea-
sonable.
The contour and three-dimensional plots presented in
Figs. 1-6 were produced for each pair of factors, whereas
the other two factors were taken as a constant at their
middle level.
Fig. 1 shows the effects of extraction duration and re-
peated extraction cycle on the determination of TC in
rapeseed. The maximum TC could be obtained with both
extraction duration and repeated extraction cycle locat-
ing in the medium levels. Both Higher duration and ex-
tended extraction cycle resulted in the decrease of TC,
which could be due to the equilibrium of TC dissolving
Tab1e 4. Analysis of variance.
Source D. F. Sum of SquaresMean of SquaresF Value Pr＞F
Linear 4 0.3153 0.0788 0.3677 0.8290
Quadratic 4 9.5318 2.3829 11.1137 0.0001**
Cross product 6 1.2609 0.2101 0.9801 0.461
Lack of fit 10 4.4821 0.4482 23.8511 0.073
Model 14 11.1080 0.7934 5.7004 0.0035**
R2=0.9116 Adj. R2=0.7193
D. F. denotes degree of freedom; **p<0.005
(X1: duration/h; X2: repeated extraction cycle; Y1: TC yield/(mg/100g))
Figure 1. Combined effect of duration and repeated extraction cycle on TC yield.
(X1: duration/h; X3:solid-solvent ratio/(g/mL); Y1: TC yield/(mg/100g))
Figure 2. Combined effects of extraction duration and solid-solvent ratio on TC yield.

L. Wang et al. / Natural Science 1 (2009) 23-29 27
into solvent obtained at 6 h and some other compounds
also extracted together with TC with further increasing
repeated extraction cycles.
Fig. 2 illustrated the effects of extraction duration and
solid-solvent ratio on the determination of TC in rape-
seed. The maximum TC was obtained with extraction
duration locating at 6h and solid-solvent ratio locating
between 1:20 and 1:30. Higher extraction duration and
solid-solvent ratio tended to result in a decrease of TC.
This also could be due to the equilibrium of TC dissolv-
ing into solvent obtained at 6 h and some other com-
pounds also extracted together with TC with further in-
creasing extraction solvent.
Fig. 3 represented the effects of extraction duration
and extraction temperature on the determination of TC in
rapeseed. The maximum TC was obtained with extrac-
tion duration locating at 6h and temperature locating
between 40°C and 45°C. Higher extraction duration and
temperature led to a decrease of TC. The reason was that
the equilibrium of TC dissolving into solvent was ob-
tained at 6 h and the cateronoids was degraded at higher
temperature.
Fig. 4 showed the effects of extraction repeated cycles
and solid-solvent ratio on the determination of TC in
rapeseed. The maximum TC was obtained with the ex-
traction repeated cycles at three times and solid-solvent
ratio with 1:30.
Fig. 5 illustrated the effects of extraction repeated cy-
cles and extraction temperature on the determination of
TC in rapeseed. The maximum TC was obtained with
extraction repeated cycles locating at three times and
extraction temperature locating between 40°C and 45°C.
Fig. 6 listed the effects of solid-solvent ratio and ex-
traction temperature on the determination of TC in rape-
seed. The maximum TC was obtained with solid- solvent
ration locating at 1:29 and extraction temperature locat-
ing between 40°C and 45°C.
From the shape of contour plots (Figs. 1-6), the in-
teraction strength as well as the optimal values range
of the independent variables could be observed.
Therefore the contour plots are generally the graphical
representation of the regression equation for the opti-
mization of extraction conditions for TC extraction
from rapeseed.
(X1: extraction duration/h; X4: extraction temperature/°C; Y1: TC yield/(mg/100g))
Figure 3. Combined effects of extraction duration and extraction temperature on TC yield.
(X2: repeated cycle;X3: solid-solvent ratio/(g/mL); Y1: TC yield/(mg/100g))
Figure 4. Combined effects of repeated cycles and solid-solvent ratio on TC yield.

28 L. Wang et al. / Natural Science 1 (2009) 23-29
(X2: repeated cycle; X4: temperature/°C; Y1: TC yield/(mg/100g))
Figure 5. Combined effects of repeated cycle and temperature on TC yield.
(X3: solid-solvent ratio/(g/mL); X4: temperature/°C; Y1: TC yield/(mg/100g))
Figure 6. Combined effects of solid-solvent ratio and temperature on TC yield.
3.3. Verification of the Model
The optimal parameters for TC extraction from rapeseed
were evaluated by RSM and a maximal TC yield of
4.791 mg/100g could be achieved at the optimal condi-
tions: extraction duration 7.3h，repeated cycles 3rd,
solid-solvent ratio 1:29, extraction temperature 42°C.
The accuracy of the model was validated with triplicate
experiments. The experimental value of TC yield was
4.77 ± 0.02mg/100g, which agreed well with the pre-
dicted value (4.79 mg/100g), which relative error be-
tween experimental value and predicted value was 0.42
± 0.04%. The verification studies proved that the pre-
dicted value of TC for the model could be realistically
achieved within a 95% confidence interval of experi-
mental values. Therefore, the model from central com-
position design was considered to be accurate and reli-
able for predicting TC yield extraction from rapeseed.
4. CONCLUSIONS
Response surface method was proved to be a powerful
tool for the optimization of extraction conditions for TC
extraction from rapeseed. The conditions were optimized
using a five-level-four-factor central composite design.
Under optimal conditions (duration 7.3h，repeated cycle
3rd, solid-solvent ratio 1:29, extraction temperature
42°C), the value of the yield of carotenoids was 4.79mg
/100g. Validation experiments were also carried out to
verify the availability and the accuracy of the model, and
the result showed that the predicted value was in well
agreement with the experimental value.
ACKNOWLEDGEMENT
The authors are very thankful for Dr. Lin Qinxiong for providing UV-
visible spectrophotometer in the experiments.
REFERENCES
[1] Holden, J. M., Eldridge, A. L. and Beeeher, G. R. (1999)
Carotenoid content of U.S. food: An update of the data-
base. Journal of Food Composition and Analysis, 12(3),
169-196.
[2] Sotirios, K. and Vassiliki, O. (2006) Antioxidant proper-
ties of natural carotenoid extracts against the AAPH-ini-
tiated oxidation of food emulsions. Innovative Food Sci-
ence and Emerging Technologies, 7(1), 132-139.

L. Wang et al. / Natural Science 1 (2009) 23-29 29
[3] Jamal, J. and Chieri, K. (2006) Variation of lycopene,
antioxidant activity, total soluble solids and weight loss
of tomato during post harvest storage. Post harvest Biol-
ogy and Technology, 41(2), 151-155.
[4] Rao, A. V. and Rao, L. G. (2007) Carotenoids and human
health. Pharmacological Research, 55(3), 207-216.
[5] Meléndez-Martínez, A. J., George, B. and Isabel, M. V.
(2007) Relationship between the colour and the chemical
structure of carotenoid pigments. Food Chemistry, 101(3),
1145-1150.
[6] Meléndez-Martínez, A. J., Vicario, I. M. and Heredia, F. J.
(2007) Review: Analysis of carotenoids in orange juice.
Journal of Food Composition and Analysis, 20(7),
638-649.
[7] Andrew, J. and Young, G. M. (2001) Antioxidant and
prooxidant properties of carotenoids. Archives of Bio-
chemistry and Biophysics, 385(1), 20-27.
[8] Adetayo, O. O. and Aluko, R. E. (2005) The anti-car-
cinogenic and anti-atherogenic effects of lycopene: A re-
view. Trends Food Sci. Technol., 16(8), 344-350.
[9] Cano, A., Acosta, M. and Arnao, M. B. (2003) Hydro-
philic and lipophilic antioxidant activity changes during
on-vine ripening of tomatoes (Lycopersicon esculentum
Mill). Post-harvest Biol. Technol., 28(1), 59-65.
[10] Akhtar, M. H. and Bryan, M. (2008) Extraction and
quantification of major carotenoids in processed foods
and supplements by liquid chromatography. Food Chem-
istry, 111(1), 255-261.
[11] Meléndez-Martínez, A. J., Vicario, I. M. and Heredia, F. J.
(2007) Provitamin A, carotenoids and ascorbic acid con-
tents of the different types of orange juices marketed in
Spain. Food Chemistry, 101(1), 177-184.
[12] Gua, Z. X., Chen, D. M., Han, Y. B., Chen, Z. G. and Gu,
F. R. (2008) Optimization of carotenoids extraction from
Rhodobacter sphaeroides. LWT-Food Science and Tech-
nology, 41(6), 1082-1088.
[13] FAO (2006) Statistical year book. Food and Agriculture
Organization of the United Nations, 2/1,2/2, Rome.
[14] Haila, H. and Heinonen, M. (1994) Action of β-carotene
on purified rapeseed oil during light storage. Food Sci-
ence and Technology, 27(6), 573-577.
[15] Hornero-Méndez, D. and Mínguez-Mosquera, M. I.
(2007) Bioaccessibility of carotenes from carrots: Effect
of cooking and addition of oil. Innovative Food Science
& Emerging Technologies, 8(3), 407-412.
[16] Rodriguez-Amaya, D. B., Kimura, M., Godoy, H. T., and
Amaya-Farfan, J. (2008) Updated Brazilian database on
food carotenoids: Factors affecting carotenoid composi-
tion. Journal of Food Composition and Analysis, 21(6),
445-463.
[17] Simkin, A. J., Moreau, H., Kuntz, M., Pagny, G., Chen-
wei, L., Tanksley, Se., and McCarthy, J. (2008) An inves-
tigation of carotenoid biosynthesis in Coffea canephora
and Coffea Arabica. Journal of Plant Physiology, 165(10),
1087-1106.
[18] Sachindra, N. M. and Mahendrakar, N. S. (2005) Process
optimization for extraction of carotenoids from shrimp
waste with vegetable oils. Bioresource Technology,
96(10), 1195-1200.
[19] Sheetal, M. C. and Laxmi (2007) Enzyme aided extrac-
tion of lycopene from tomato tissues. Food Chemistry,
102(1), 77-81.
[20] Huang, W., Li, Z. S., Niu, H., Li, D. and Zhang, J. (2008)
Optimization of operating parameters for supercritical
carbon dioxide extraction of lycopene by response sur-
face methodology. Journal of Food Engineering, 89(3),
298-302.
[21] Mateose, R. and Garcia-Mesa, J. A. (2006) Rapid and
quantitative extraction method for the determination of
chlorophylls and carotenoids in olive oil by
high-performance liquid chromatography. Analytical and
Bioanalytical chemistry, 385(7), 1247-1254.
[22] Rodriguez-Amaya, D. B. (2001) A guide to carotenoid
analysis in foods. Washington, DC.: Ed. ILSI-Interna-
tional Life Sciences Institute.
[23] Schoefs, B. (2002) Chlorophyll and carotenoid analysis
in food products. Properties of the pigments and methods
of analysis. Trends in Food Science and Technolog.,
13(11), 361-371.