Available online at www.sciencedirect.

com

Separation and Purification Technology 61 (2008) 358–365

Supercritical carbon dioxide extraction of rice bran oil

and column partition fractionation of ␥-oryzanols
Chao-Rui Chen a , Ling-Ya Wang a , Chih-Hung Wang a , Wai-Jane Ho b , Chieh-Ming J. Chang a,∗
a Department of Chemical Engineering, National Chung Hsing University, #250, KuoKuang Road, Taichung 402, Taiwan, ROC
b Department of Bioresources, Dayen University, #112, ShanJiao Road, ChangHua 515, Taiwan, ROC

Received 8 August 2007; received in revised form 29 October 2007; accepted 1 November 2007

Abstract
This work studies how pressure and temperature affect supercritical carbon dioxide extraction of rice bran oil from powdered rice bran, followed
by the concentration and isolation of ␥-oryzanols by column partition purification. The two purest ␥-oryzanols (>98 wt.%) were isolated from
reagent chemical ␥-oryzanols by preparative reverse-phase high-pressure liquid chromatography. In searching for a suitable range of extraction
conditions, supercritical extraction at 350 bar and 313 K yielded 17.5% oil and the extraction efficiency of ␥-oryzanols was 84.9%, using 1200 g of
carbon dioxide over 4 h. Finally, supercritical extractions at pressures from 250 to 350 bar and at temperatures from 313 to 333 K were chosen based
on the response surface methodology to determine their effects on concentration of ␥-oryzanols in the extracted oil. Pressure proved more significant
than temperature in increasing the concentration of ␥-oryzanols. A significant amount of the purest oil that contained 37.0 wt.% ␥-oryzanols was
obtained following a normal-phase medium-pressure column partition fractionation of the extracted oil.
© 2007 Elsevier B.V. All rights reserved.

Keywords: Column chromatography; Extraction; ␥-Oryzanols; Rice bran; Supercritical carbon dioxide

1. Introduction concluded that cycloartenyl ferulate, 24-methylenecycloartanyl

ferulate and campesteryl ferulate are three major ␥-oryzanols
Rice bran is comprised of 11–15% proteins, 34–62% carbo- in rice bran. These ␥-oryzanols have several biological and
hydrates, 7–11% crude fibers, 7–10% ashes and 15–20% lipids, physiological effects, such as serving as anti-oxidation [5],
which are by-products of the rice-refining process [1]. This anti-blood cholesterol [6] and anti-carcinogenic agents [7–10].
material contains nutritional and non-nutritional compounds that The LC–MS/MS method is frequently adopted in the structural
benefit humans. Rice bran oil contains 95.6% saponifiable lipids, identification of ␥-oryzanols [11]. Stöggl et al. simultane-
such as glycolipid and phospholipids and 4.2% unsaponifiable ously identified and quantified tocopherols, ␥-oryzanols and
lipids, such as tocopherols, tocotrienols, ␥-oryzanols, sterols cartenoids [12]. Aguilar-Garcia et al. also performed a bio-
and cartenoids [2]. Lipids are comprised of mostly triglycerides. logical study of a correlation between ␥-oryzanols content and
However, a lipase can hydrolyze these triglycerides to yield free anti-oxidant capacity [13].
fatty acids [3]. Wastes of edible rice bran oil can be processed into bio-
The content of ␥-oryzanols in the extracted rice bran oil is diesel by acid-catalyzed transesterification [14]. In recent years,
approximately 1.8–3%, according to Hu et al. [3]. Xu and God- supercritical fluid extractions of powdered rice bran have
ber [4] employed a low-pressure normal-phase silica column to demonstrated that the amount of extracted oil is equivalent to
yield an enrichment of oryzanol in the original oil. This RBO that obtained by Soxhlet N-hexane extraction [15]. Deacidi-
was further partitioned using a preparative normal-phase HPLC fication of rice bran oil using supercritical carbon dioxide is
method. Finally, ten ␥-oryzanols have been identified by reverse- recognized as a very environmental friendly process. Kim et
phase HPLC coupled with GC-Mass chromatography. They al. investigated the time-related mass transfer kinetics of oil
components between solid-phase (rice bran) and fluid-phase
(carbon dioxide) in supercritical carbon dioxide extraction [16].
∗ Corresponding author. Tel.: +886 4 2285 2592; fax: +886 4 2286 0231. Chang et al. studied supercritical carbon dioxide extraction
E-mail address: cmchang@dragon.nchu.edu.tw (C.-M.J. Chang). kinetics and high-pressure vapor–liquid-phase equilibrium mea-

1383-5866/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.seppur.2007.11.003
 C.-R. Chen et al. / Separation and Purification Technology 61 (2008) 358–365 359

surements between oil compounds and carbon dioxide [17,18]. tion system. Analytical grade reagents—99.95% carbon dioxide
Many multi-stage supercritical fluid deacidifications of rice bran (Toyo gas, Taiwan), 99.7% nitrogen (Toyo gas), 99.8% hydrogen
oil have been conducted to remove free fatty acids or triglyc- (Toyo gas), 99% air (Toyo gas), 99.9% ethyl acetate (Mallinck-
erides from raffinate oil using a packed column at a middle–high rodt, USA), 99.9% N-hexane (Mallinckrodt), 99.9% methanol
pressure at 333–353 K to increase concentrations of ␥-oryzanols (Mallinckrodt), 99.9% dichloromethane (Mallinckrodt), 99.5%
and phytosterols in the oil [19–24]. ethyl ether (Mallinckrodt), 99.5% isopropanol (Mallinckrodt),
Except for cycloartenyl ferulate, no pure standard of ␥- 99.9% acetonitrile (J.T. Baker, USA), 99.8% acetic acid (Merck,
oryzanols can be obtained from local chemical suppliers. This Germany) and 95% ethanol (Taiwan Sugar Co. Ltd., Taiwan),
goal of this work was to obtain pure 24-methylenecycloartanyl were purchased from a local supplier and used without further
ferulate and campesteryl ferulate from reagent chemical ␥- treatment. Authentic standards including 98% cycloartenyl fer-
oryzanols using a semi-preparative HPLC method. Supercritical ulate (Wako, Japan), 95% mixed standard ␥-oryzanols (Wako)
carbon dioxide extraction of rice bran oil from powdered rice and a fatty acid methyl esters standard (Supelco, USA) were
bran, designed using response surface methodology, followed used herein only for quantification.
by the concentration of ␥-oryzanols in the extracted oil using a
middle-pressure column partition fractionation, was studied. 2.2. Isolation and identification of two γ-oryzanols

2. Experimental method A semi-preparative HPLC system using one C18 column

(YMC-C18, 250 mm × 10 mm i.d.) that is connected to a UV
2.1. Reagents and materials detector (PE-785A, PerkinElmer) via a high-pressure pump (PE-
410, PerkinElmer). The 1.5 mL (500 ppm) mixed standard of
Ten kilograms of fresh rice bran powder were donated by ␥-oryzanols dissolved in methanol was injected into a 5 mL
a local Da-quan rice bran producer (Taichung, Taiwan) and loop and this sample was partitioned using a mixed solvent of
were stored in a cooler at 269 K before use. The water content acetonitrile, dichrolomethane and acetic acid (90:6:4) at a flow
of this rice bran powder was to be under 10%, as determined rate of 5 mL/min. The eluates were detected at a wavelength
using a moisture meter (A&D, AD-4714A, Japan). De-ionized of 330 nm and four isolated compounds were collected based
water was obtained using a Milli-Q reverse osmosis purifica- on their retention time. The isolated 24-methylenecycloartanyl

Fig. 1. Schematic flow diagram of SC-CO2 extraction of rice bran oil.

360 C.-R. Chen et al. / Separation and Purification Technology 61 (2008) 358–365

ferulate and campesteryl ferulate were obtained at the 10 mg ISCO, USA) (5), and was heated to supercritical conditions
level. Their chemical structures were identified using a 400 Hz using a double-pipe heat exchanger (8) and a re-boiler (10-
1 H NMR spectrophotometer. 2). This carbon dioxide, maintained at a flow rate of 5 L/min
(STP), flowed into the extractor (10-1); came into contact with
2.3. Quantification of γ-oryzanols, free fatty acids and the rice bran powder, and was used to extract the oil. A heat-
triglycerides ing element equipped with a PID temperature controller (4)
was thermostatically maintained at a constant temperature; two
HPLC quantifications of four ␥-oryzanols (cycloartenyl fer- back-pressure regulators (12-1, 12-2) located at the outlet were
ulate, 24-methylenecycloartanyl ferulate, campesteryl ferulate manually adjusted to maintain constant extraction pressure. Fol-
and sitosteryl ferulate) and three free fatty acids (oleic acid, lowing the extraction, CO2 entered into a 130 mL separator (14)
linoleic acid and linolenic acid) were performed using a reverse- by a drop in pressure and was expanded through a spiral-type
phase analytical column (YMC-C18, 5 ␮m, 250 mm × 4.6 mm nozzle. The SC-CO2 extract then passed through a heated line
i.d., Japan). The column was linked to a UV/vis detector (L- connector (17) and entered in a separator that was maintained
4200H, Hitachi, Japan) by a high-pressure pump (L-7100, at 50 bar and 303 K. The amount of CO2 used was measured
Hitachi, Japan) and the software to control the interface (D- using a wet gas meter (W-NK-Da-1B, Shinagawa, Japan) (15)
7000, Hitachi, Japan). The column temperature was maintained and thus returned to the ambient conditions. Precipitates (15%
at 313 K using a column oven (TCM-004657, Waters-Millipore, extracted oil) that had washed out from the back-pressure regu-
USA). The absorption of UV from the samples was detected lators and the transfer lines were then mixed with the collected
at a wavelength of 330 and 210 nm for the each of the four ␥- oil (85% oil) to prepare them for analysis. At the end of each
oryzanols and the three free fatty acids. The injection volume of experiment, the extracted solution was collected and the solvent
the sample was 20 ␮L. One mobile phase of the mixed solvent was dried out using a vacuum rotary evaporator. The residue was
of 90% acetonitrile, 6% dichrolomethane and 4% acetic acid then weighed and stored at 273 K before use.
by weight fraction was used in the ␥-oryzanols analysis. The
other mobile phase of the mixed solvent of 85% acetonitrile, 2.6. Column partition fractionation of γ-oryzanols oil
5% methanol and 10% de-ionized water with 1% acetic acid
was used to analyze the free fatty acids. Both R-square corre- Solid–liquid column purification, a form of partition extrac-
lation coefficients of the calibration curves of ␥-oryzanols and tion, has been used in the isolation of bioactive compounds from
free fatty acids were at least 0.999. extracted oil [25–27]. In this study, a total 15 g of SC-CO2 -
GC quantification of seven triglycerides was performed using extracted oil was obtained and diluted with N-hexane in a batch.
a non-polar capillary column (DB-5, J&W, USA) in a gas The diluted solution was then loaded into a medium-pressure
chromatograph (GC-14B, Shimadzu, Japan). The column tem- column (440 mm L × 37 mm i.d.) that was packed with silica
perature was set to 443 K initially, and programmed to increase to gel 60 resin (LiChroprep-Si-60, 40–63 ␮m, Merck, Germany).
488 K at 5 K/min, then to 496 K at 2 K/min, and finally to 503 K The diluted solution of SC-CO2 -extracted oil was pumped into
at 1 K/min. The split ratio was 3.4: 1, and the flame ionization the medium-pressure column through an intelligent pump (PU-
detector temperature was set to 553 K.

2.4. Classical extraction of solvents

For the Soxhlet solvent extraction, 15 g of rice bran pow-

der was loaded into a 50 mL Soxhlet timber-type extractor with
reflux having 16 turnover numbers and extracted using 300 mL
N-hexane, isopropanol and methanol individually for 4 h. All of
the extracts were collected and weighed. The total amount of
extract, and the extraction efficiencies of ␥-oryzanols, free fatty
acids and triglycerides were then calculated.

2.5. Supercritical carbon dioxide extraction of rice bran oil

Fig. 1 displays a schematic flow diagram of SC-CO2 extrac-

tion from rice bran. 10 g was rice bran powder was packed inside
a 250 mL stainless steel tubular extractor with an internal diam-
eter of 2.5 cm and a height of 54 cm. A specified amount of
glass wool was packed into the two ends of the extractor to pre-
vent the rice bran powder escaping from the extractor. Liquid
CO2 from a cylinder (1) into which was inserted a siphon-tube
passed through a cooling bath (3) at 277 K, was compressed to Fig. 2. Block diagram of solid–liquid column partition fractionation of rice bran
the desired working pressure using a syringe pump (100DX, oil.
Table 1
Experimental data on Soxhlet extractions and SC-CO2 extractions of rice bran oil
Exp# WCO2 (g) Woil (g) TY (%) WOry ∗
WOry ROry βOry WFFA ∗
WFFA RFFA βFFA WTG ∗
WTG RTG βTG Others
(mg/goil ) (mg/gRB ) (%) (mg/goil ) (mg/gRB ) (%) (mg/goil ) (mg/gRB ) (%) (mg/goil )

Soxhleta (15 g rice bran)

1 – 3.00 20.0 15.2 3.03 100.0 5.00 95.0 19.0 100.0 5.00 800 160.0 100.0 5.00 89.8
2 – 3.97 26.4 9.07 2.40 79.1 2.99 68.1 18.0 94.8 3.59 555 146.7 91.7 3.47 367.9
3 – 3.49 23.3 12.3 2.87 94.5 4.06 70.0 16.3 85.7 3.68 574 133.6 83.5 3.59 343.7
SC-CO2 b (10–20 g rice bran)
4 600 1.64 16.4 13.9 2.28 75.2 4.59 101.0 16.5 87.0 5.31 793 129.8 81.1 4.95 92.3
5 800 1.69 16.9 13.5 2.28 75.3 4.46 102.0 17.2 90.6 5.37 793 133.9 83.7 4.96 91.5

C.-R. Chen et al. / Separation and Purification Technology 61 (2008) 358–365

6 1200 1.75 17.5 15.2 2.67 87.9 5.01 101.4 17.8 93.5 5.33 798 139.8 87.4 4.99 85.7
7 1200 2.65 17.6 14.6 2.57 84.8 4.81 105.6 18.6 98.0 5.56 787 138.8 86.7 4.92 92.9
8 1600 2.84 14.2 10.3 1.46 48.1 3.39 92.9 13.2 69.4 4.89 789 112.0 70.0 4.93 107.6

(1) 342 K N-hexane; (2) 356 K isopropanol; (3) 338 K methanol; WCO2 : consumption of carbon dioxide; Woil : weight of the extracted; TY: total oil yield = (Woil /WRB ) × 100%; WRB : weight of rice bran;
WOry : weight of oryzanols; WOry ∗ : yield of oryzanols; W ∗ ∗
FFA : weight of free fatty acids; WFFA : yield of free fatty acids; WTG : weight of triglycerides; WTG : yield of triglycerides; ROry : oryzanols extraction
efficiency = (WOry /WOry,Soxhlet ) × 100%; RFFA : free fatty acids extraction efficiency = (WFFA /WFFA,Soxhlet ) × 100%; RTG : triglycerides extraction efficiency = (WTG /WTG,Soxhlet ) × 100%; βOry : concentration factor of
oryzanols = ROry /TY; βFFA : concentration factor of free fatty acids = RFFA /TY; βTG : concentration factor of triglycerides = RTG /TY; others: weight of waxes, glycolipids and phospholipids; TYSoxhlet = 20.0 (%);
∗
WOry,Soxhlet ∗
= 3.03 (mg/g); WFFA,Soxhlet ∗
= 19.0 (mg/g); WTG,Soxhlet = 160.0 (mg/g); βOry,FFA,TG,Soxhlet = 5.00.
a Four hours Soxhlet extraction.
b SC-CO extractions at 350 bar, 313 K, 10 g/min.
2

Table 2
Response surface-designed SC-CO2 extractions of 10 g of rice bran powder
RSM# P (bar) T (K) Woil TY WOry ∗
WOry ROry βOry WFFA ∗
WFFA RFFA βFFA WTG ∗
WTG RTG βTG Others
(g) (%) (mg/goil ) (mg/gRB ) (%) (mg/goil ) (mg/gRB ) (%) (mg/goil ) (mg/gRB ) (%) (mg/goil )

1(A) 250 313 0.86 8.6 4.6 0.40 13.1 1.51 170 14.7 77.3 8.96 784 67.6 42.3 4.90 41.2
2(F) 250 323 0.82 8.2 4.8 0.39 12.8 1.57 175 14.3 75.4 9.20 776 63.6 39.7 4.85 44.2
3(A) 250 333 0.82 8.2 4.7 0.38 12.6 1.53 217 17.8 93.8 11.42 721 59.3 37.1 4.51 57.0
4(F) 300 313 1.48 14.8 9.0 1.32 43.7 2.96 119 17.6 92.6 6.27 800 118.2 73.9 5.00 71.6
5(C) 300 323 1.43 14.3 6.5 0.92 30.5 2.14 121 17.2 90.6 6.36 786 112.0 70.0 4.91 86.6
6(F) 300 333 1.52 15.2 7.5 1.14 37.6 2.48 117 17.7 93.2 6.15 792 120.1 75.0 4.95 83.0
7(A) 350 313 1.75 17.5 14.7 2.57 84.9 4.86 105 18.4 96.7 5.54 798 139.3 87.1 4.99 82.2
8(F) 350 323 1.73 17.3 13.9 2.40 79.3 4.59 104 18.0 94.6 5.48 760 131.2 82.0 4.75 122.2
9(A) 350 333 1.73 17.3 14.7 2.55 84.1 4.86 104 18.0 94.8 5.47 747 129.4 80.8 4.67 134.4

P: pressure; T: temperature; Woil : weight of the extracted; TY: total oil yield = (Woil /WRB ) × 100%; WRB : weight of rice bran; WOry : weight of oryzanols; WOry ∗ : yield of oryzanols; W
FFA : weight of free
∗ : yield of free fatty acids; W : weight of triglycerides; W ∗ : yield of triglycerides; R
fatty acids; WFFA : oryzanols extraction efficiency = (W /W ) × 100%; R : free fatty acids extraction effi-
TG TG Ory Ory Ory,Soxhlet FFA
ciency = (WFFA /WFFA,Soxhlet ) × 100%; RTG : triglycerides extraction efficiency = (WTG /WTG,Soxhlet ) × 100%; βOry : concentration factor of oryzanols = ROry /TY; βFFA : concentration factor of free fatty acids = RFFA /TY;
∗
βTG : concentration factor of triglycerides = RTG /TY; others: weight of waxes, glycolipids and phospholipids; TYSoxhlet = 20.0 (%); WOry,Soxhlet ∗
= 3.03 (mg/g); WFFA,Soxhlet ∗
= 19.0 (mg/g); WTG,Soxhlet = 160.0 (mg/g);
βOry,FFA,TG,Soxhlet = 5.00.

361
362 C.-R. Chen et al. / Separation and Purification Technology 61 (2008) 358–365

1580, Jasco, Japan). Chromatography was then employed to were obtained at a CO2 to solids ratio of 120. The amount of oil
fractionate the ␥-oryzanols with an affinity for fat from other extracted declined as the amount of solids increased (datum# 8
species in the oil. The eluting solvent consisted of N-hexane (H) in Table 1).
and ethyl acetate (EA).
Fig. 2 presents an information block diagram. In these 3.3. Experimental designed SC-CO2 extraction
stepwise elution processes, most triglycerides and free fatty
acids were initially removed by the eluent but the ␥-oryzanols A two-factor parameter (pressure and temperature) response
remained adsorbed and were later eluted. The flow rate to the col- surface methodology (RSM), with four axial points, four factor
umn was 2 mL/min. Five eluting solvent mixtures of N-hexane
with ethyl acetate in volume ratios of 100:0, 98:2, 95:5, 90:10
and 0:100 were used in this fractionation study. After the solvent
had been dried out, each fraction was concentrated and analyzed.

3. Results and discussion

3.1. Two purified components of γ-oryzanols

The 24-methylenecycloartanyl ferulate chemical shifts

pattern of 1 H NMR in CDCl3 , δ was 7.60 (d, J = 15.8 Hz), 7.08
(dd, J = 1.8, 8.1 Hz), 7.04 (d, J = 1.8 Hz), 6.92 (d, J = 8.4 Hz),
6.30 (d, J = 16.1 Hz), 5.85 (br s), 4.72 (br s), 4.71 (m), 4.67
(br s), 3.94 (s), 2.24 (sept., J = 7.0 Hz), 1.04 (d, J = 7.0 Hz),
1.03 (d, J = 6.6 Hz), 0.98 (s), 0.92 (s), 0.9 (d, J = 5.5 Hz), 0.9
(s), 0.60 (d, J = 4.0 Hz, endo), 0.37 (d, J = 4.0 Hz, exo). The
campesteryl ferulate chemical shifts pattern of 1 H NMR in
CDCl3 , δ was 7.60 (d, J = 15.8 Hz), 7.07 (dd, J = 1.8, 8.4 Hz),
7.03 (d, J = 1.8 Hz), 6.91 (d, J = 8.4 Hz), 6.28 (d, J = 16.1 Hz),
5.84 (br s), 5.40 (m), 4.71 (m), 3.92 (s), 2.41 (br s), 2.39 (br
s), 1.05 (s), 0.92 (d, J = 6.6 Hz), 0.85 (d, J = 6.2 Hz), 0.81 (d,
J = 6.8 Hz), 0.78 (d, J = 6.6 Hz), 0.69 (s). Yasukawa et al. had
already described these spectra patterns [7].

3.2. Extractions of rice bran oil

Table 1 presents experimental data on a few Soxhlet and SC-

CO2 extractions of rice bran oil. The maximum concentration
of ␥-oryzanols in the N-hexane-extracted oil was 15.2 mg/goil ,
which is the ␥-oryzanols content in rice bran powder. A concen-
tration of 3.03 mg/gpowder was obtained in 4 hr from a 300 mL
Soxhlet N-hexane extraction of 15 g rice bran powder. This rep-
resents 100% recovery of ␥-oryzanols from the rice bran powder.
The maximal amount of this N-hexane-extracted oil was 3 g,
which was highest among all of these organic solvent extrac-
tions. The amount of SC-CO2 -extracted oil is less than that of
Soxhlet N-hexane-extracted oil. These two constituents reached
90%, as indicated in Table 1.
Parameters of SC-CO2 extraction that potentially influence
the extraction efficiency include solids in the feed, consumption
of carbon dioxide, the temperature and the pressure. A set of
preliminary experimental tests were performed amounts of rice
bran powder ranged from 10 to 20 g; amounts of CO2 consumed
ranged from 600 to 1600 g. The experimental data obtained in
these tests revealed that the amount of solids and the consump-
tion of carbon dioxide markedly affect the extraction efficiency Fig. 3. Effects of pressure and temperature on SC-CO2 extractions of rice bran
of rice bran oil. The extracted amount of oil and the extrac- oil, (1) total oil yield; (2) extraction efficiency of oryzanols; (3) concentration fac-
tion efficiency of ␥-oryzanols increased with the carbon dioxide tor of oryzanols in 3D. (F-testing: R2(1) = 0.9979, S.D.(1) = 0.30; R2(1) = 0.9943,
consumption; 17.5% of oil and 87.9% recovery of ␥-oryzanols S.D.(2) = 3.83; R2(3) = 0.9874, S.D.(3) = 0.27).
 C.-R. Chen et al. / Separation and Purification Technology 61 (2008) 358–365 363

points and one center point was designed for the SC-CO2
extraction of rice bran oil from the 10 g of rice bran powder.
The flow rate of CO2 was 10 g/min and the amount of CO2
consumed was 1200 g, as determined from the results in Table 1.
Table 2 presents experimental data on RSM SC-CO2 extraction
at pressures from 250 to 350 bar and temperatures from 313 to
333 K.
Fig. 3 shows that the total oil yield (TY), the extraction effi-
ciency (ROry ) and the concentration factor (βOry ) reaches the
maxima of 17.5%, 84.9% and 4.86, respectively, at 350 bar
and 313 K (datum# 7 in Table 2). The pressure effect seems
to be more significant than the temperature effect. The total
oil yield, extraction efficiency of ␥-oryzanols (ROry ) and con-
centration factor of ␥-oryzanols (βOry ) increased with pressure,
but decreased as the temperature increased. Fig. 4 reveals that
the extraction efficiency of free fatty acids (RFFA ) and the con-
centration factor (βFFA ) reached minima of 10.4% and 5.47,
respectively, at 350 bar and 333 K (datum# 9 in Table 2). The
amount of free fatty acids in extracted oil and the concentra-

Fig. 5. Effects of pressure and temperature on SC-CO2 extractions of rice

bran oil, (1) extraction efficiency of triglycerides; (2) concentration factor of
triglycerides in 3D. (F-testing: R2(1) = 0.9930, S.D.(1) = 2.72; R2(2) = 0.9830,
S.D.(2) = 0.07).

tion of free fatty acids both decreased as the pressure increased,

and the temperature did not influence either. Fig. 5 shows the
three-dimensional extraction efficiency of triglycerides, which
reached a maximum of 87.1% at 350 bar and 313 K extraction
(datum# 7 in Table 2). The concentration factor of triglycerides
(βTG ) reaches 5.0 at 300 bar and 313 K extraction (datum# 4 in
Table 2). The extraction efficiency (RTG ) and the concentration
of triglycerides also increased with pressure.

3.4. Deacidification of rice bran oil using column partition

fractionation

A medium-pressure normal-phase column partition fraction-

ation was adopted herein to purify ␥-oryzanols from the Soxhlet
and SC-CO2 -extracted oil. Table 3 presents the amount of ␥-
oryzanols, triglycerides, free fatty acids and total fatty acids in
Fig. 4. Effects of pressure and temperature on SC-CO2 extractions of rice bran
oil, (1) extraction efficiency of free fatty acids; (2) concentration factor of each fraction. After the sixth fraction of N-hexane oil (datum#
free fatty acids in 3D. (F-testing: R2(1) = 0.9705, S.D.(1) = 1.14; R2(2) = 0.9703, F6 in Table 3) was eluted using the 90:10 solution of N-hexane
S.D.(2) = 0.60). and ethyl acetate, 37.0 wt.% of ␥-oryzanols was obtained. The
364 C.-R. Chen et al. / Separation and Purification Technology 61 (2008) 358–365

Table 3
Experimental data on column partition fractionation of rice bran oil
Sample Eluent Woil (g) Ory (%) TG (%) FFA (%) FA (%) Others (%)

Soxhlet N-hexane extracted oil

Feed F 20.2 2.1 53.9 40.0 93.9 4.0
F1 100:0 0.91 0 42.9 0 42.9 57.1
F2 200:1 6.61 0 61.5 30.5 92.0 8.0
F3 98:2 2.94 0 43.6 52.3 95.9 4.1
F4 95:5 2.19 0.6 30.8 64.3 95.1 4.3
F5 90:10 0.72 4.4 28.9 65.0 93.9 1.7
F6 90:10 0.92 37.0 22.5 32.5 55.0 8.0
F7 0:100 1.90 0.2 63.0 25.1 88.1 11.7
SC-CO2 extracted oil
Feed J 14.7 1.3 83.6 12.0 95.6 3.1
J1 100:0 8.99 0 96.9 0 96.9 3.1
J2 98:2 0.30 0 93.4 0 93.4 6.6
J3 95:5 1.63 0.3 71.2 23.2 94.4 5.3
J4 90:10 0.55 0.6 45.6 51.4 97.0 2.4
J5 90:10 0.84 18.9 48.1 23.6 71.7 9.4
J6 0:100 1.41 0.6 81.4 11.6 93.0 6.4

Eluent: N-hexane + ethyl acetate; Ory: percentage of oryzanols = (WOry /Woil ) × 100%; TG: percentage of triglycerides = (WTG /Woil ) × 100%; FFA: percentage
of free fatty acids = (WFFA /Woil ) × 100%; FA: percentage of fatty acids = (WFA /Woil ) × 100%; Others: percentage of waxes, glycolipids and phospho-
lipids = (Woil − WOry − WFA )/Woil × 100%; Woil : weight of the extracted; WOry : weight of oryzanols; WTG : weight of triglycerides; WFFA : weight of free fatty
acids; WFA : weight of fatty acids = WTG + WFFA .
Bold values represent the best values.

fifth fraction that contained 18.9 wt.% of ␥-oryzanols of the SC- extracted oil. Separating free fatty acids and triglycerides from
CO2 -extracted oil was obtained using the same mobile solvent. the purified ␥-oryzanols fraction remains difficult, and requires
The concentration of free fatty acids in the N-hexane-extracted further study.
oil from rice bran powder containing 40% FFA was higher than
that in the SC-CO2 -extracted oil from the powder with 10% FFA, Acknowledgements
so the purity of ␥-oryzanols was lower. In these column parti-
tion fractionations, most triglycerides and free fatty acids had The authors would like to thank the National Science Council
been removed by preceding solvent elutions of large fractions of of the Republic of China, Taiwan, for financially supporting this
N-hexane. However, the remaining elutions retained a few free research under Contract No. NSC95-2221-E005-133. This work
fatty acids in the purified ␥-oryzanols fraction, as revealed by is also supported in part by the ministry of education, Taiwan,
data# F6 and J5 in Table 3. ROC under the ATU plan.

4. Conclusions References

This study investigated the supercritical carbon dioxide [1] B.O. Juliano, P.A. Hicks, Rice functional properties and rice food products,
Food Rev. Int. 12 (1996) 71–103.
extraction of ␥-oryzanols-containing rice bran oil from pow- [2] R.N. Sayre, Rice bran as a source of edible oil and higher value chemicals,
dered rice bran using a design based on the response surface Western Regional Research Center, ARS, USDA (1988).
methodology and studied the fractionation of ␥-oryzanols [3] W. Hu, J.H. Wells, T.S. Shin, J.S. Godber, Comparison of isopropanol and
using medium-pressure column partition chromatography. Two hexane for extraction of vitamin E and oryzanols from stabilized rice bran,
purest ␥-oryzanols were successfully isolated from a mix- J. Am. Oil Chem. Soc. 73 (1996) 1653–1656.
[4] Z. Xu, J.S. Godber, Purification and identification of components of ␥-
ture of ␥-oryzanols using preparative C18 high-pressure liquid oryzanols in rice bran oil, J. Agric. Food Chem. 47 (1999) 2724–2728.
chromatography. Experimental supercritical carbon dioxide [5] J.K. Duve, P.J. White, Extraction and identification of antioxidants in oats,
extraction indicated that the pressure dominated the increase in J. Am. Oil Chem. Soc. 68 (1991) 365–370.
the concentration of ␥-oryzanols in the extracted oil. The con- [6] G.S. Seetharamaiah, N. Chandrasekhara, Studies on hypocholesterolemic
centration of ␥-oryzanols in the SC-CO2 oil was of the same activity of rice bran oil, Atherosclerosis 78 (1989) 219–223.
[7] K. Yasukawa, T. Akihisa, Y. Kimura, T. Tamura, M. Takido, Inhibitory
order of magnitude as that in the Soxhlet N-hexane oil. How- effect of cycloartenol ferulate, a component of rice bran, on tumor pro-
ever, the light-yellowish color, odor and simple post-removing motion in two-stage carcinogenesis in mouse skin, Biol. Pharm. Bull. 21
of paraffins from the SC-CO2 extracted oil is superior to that (1998) 1072–1076.
obtained by Soxhlet N-hexane extraction. Again, the SC-CO2 [8] S.H. Nam, S.P. Choi, M.Y. Kang, N. Kozukue, M. Friedman, Antioxida-
tive, antimutagenic, and anticarcinogenic activities of rice bran extracts in
extraction has been recognized as an environmental friendly
chemical and cell assays, J. Agric. Food Chem. 53 (2005) 816–822.
method to produce edible oil. A purified fraction that con- [9] T. Akihisa, K. Yasukawa, M. Yamaura, M. Ukiya, Y. Kimura, N. Shimizu,
tained 18.9% of ␥-oryzanols was obtained using a normal-phase K. Arai, Triterpene alcohol and sterol ferulates from rice bran and their
medium-pressure column partition fractionation of the SC-CO2 - anti-inflammatory effects, J. Agric. Food Chem. 48 (2000) 2313–2319.
 C.-R. Chen et al. / Separation and Purification Technology 61 (2008) 358–365 365

[10] H.F. Luo, Q. Li, S. Yu, T.M. Badger, N. Fang, Cytotoxic hydroxylated [19] Z. Shen, M.V. Palmer, S.S.T. Ting, R.J. Fairclough, Pilot scale extraction
triterpene alcohol ferulates from rice bran, J. Nat. Products 68 (2005) 94–97. and fractionation of rice bran oil using supercritical carbon dioxide, J.
[11] N. Fang, S. Yu, T.M. Badger, Characterization of triterpene alcohol and Agric. Food Chem. 45 (1997) 4540–4544.
sterol ferulates in rice bran using LC–MS/MS, J. Agric. Food Chem. 51 [20] N.T. Dunford, J.W. King, Phytosterol enrichment of rice bran oil by a
(2003) 3260–3267. supercritical carbon dioxide fractionation technique, J. Food Sci. 65 (2000)
[12] W. Stöggl, C. Huck, S. Wongyai, H. Scherz, G. Bonn, Simultaneous deter- 1395–1399.
mination of carotenoids, tocopherols, and ␥-oryzanols in crude rice bran oil [21] N.T. Dunford, J.W. King, Thermal gradient deacidification of crude rice
by liquid chromatography coupled to diode array and mass spectrometric bran oil utilizing supercritical carbon dioxide, J. Am. Oil Chem. Soc. 78
detection employing silica C30 stationary phases, J. Sep. Sci. 28 (2005) (2001) 121–125.
1712–1718. [22] J.W. King, N.T. Dunford, Phytosterol-enriched triglyceride fractions from
[13] C. Aguilar-Garcia, G. Gavino, M. Baragaño-Mosqueda, P. Hevia, V.C. vegetable oil deodorizer distillates utilizing supercritical fluid fractionation
Gavino, Correlation of tocopherol, tocotrienol, ␥-oryzanols and total technology, Sep. Sci. Technol. 37 (2002) 451–462.
polyphenol content in rice bran with different antioxidant capacity assays, [23] N.T. Dunford, J.A. Teel, J.W. King, A continuous countercurrent supercrit-
Food Chem. 102 (2007) 1228–1232. ical fluid deacidification process for phytosterol ester fortification in rice
[14] S. Zullaikah, C.C. Lai, S.R. Vali, Y.H. Ju, A two-step acid-catalyzed process bran oil, Food Res. Int. 36 (2003) 175–181.
for the production of biodiesel from rice bran oil, Bioresour. Technol. 96 [24] L. Danielski, C. Zetzl, H. Hense, G. Brunner, A process line for the pro-
(2005) 1889–1896. duction of raffinated rice oil from rice bran, J. Supercrit. Fluids 34 (2005)
[15] Z. Shen, M.V. Palmer, S.S.T. Ting, R.J. Fairclough, Pilot scale extraction 133–141.
of rice bran oil with dense carbon dioxide, J. Agric. Food Chem. 44 (1996) [25] Y.N. Lee, C.R. Chen, H.L. Yang, C.C. Lin, C.J. Chang, Isolation and purifi-
3033–3039. cation of 3,5-diprenyl-4-hydroxycinnamic acid (artepillin C) in Brazilian
[16] H.J. Kim, S.B. Lee, K.A. Park, I.K. Hong, Characterization of extraction propolis by supercritical fluid extractions, Sep. Purif. Technol. 54 (2007)
and separation of rice bran oil rich in EFA using SFE process, Sep. Purif. 130–138.
Technol. 15 (1999) 1–8. [26] L.H. Chang, C.J. Chang, Continuous hot pressurized fluids extraction of
[17] C.J. Chang, C.C. Chen, High-pressure densities and P-T–x-y diagrams isoflavones and soyasaponins from defatted soybean flakes, J. Chin. Inst.
for carbon dioxide + linalool and carbon dioxide + limonene, Fluid Phase Chem. Eng. 38 (2007) 313–319.
Equilibr. 163 (1999) 119–126. [27] C.R. Chen, M.R. Lee, Y.N. Lee, I.C. Wei, C.M.J. Chang, Hot pressurized
[18] C.J. Chang, M.S. Lee, B.C. Li, P.Y. Chen, Vapor–liquid equilibria of CO2 fluid extraction of flavonoids and phenolic acids from Brazilian propolis
with four unsaturated fatty acid esters at elevated pressure, Fluid Phase and their cytotoxic assay in-vitro, J. Chin. Inst. Chem. Eng. 38 (2007)
Equilibr. 233 (2005) 56–65. 191–196.