ARTICLE IN PRESS

http://www.paper.edu.cn
Biosystems Engineering (2006) 93 (4), 383–391
doi:10.1016/j.biosystemseng.2006.01.008
PH—Postharvest Technology

Molecular Distillation for recovering Tocopherol and Fatty Acid Methyl Esters
from Rapeseed Oil Deodoriser Distillate
S.T. Jiang1,2; P. Shao1; L.J. Pan1,2; Y.Y. Zhao1,2
1
School of Biotechnology and Food Engineering, Hefei University of Technology, Hefei 230009, China; e-mail of corresponding author:
[email protected]
2
The Key Laboratory of Agricultural Production Biochemistry of the Ministry of Education, Hefei 230069, China

(Received 28 February 2005; accepted in revised form 10 January 2006; published online 9 March 2006)

Tocopherol and fatty acid methyl esters (FAME) were recovered through a process involving acid-catalysed
methyl esterification and crystallisation of sterols followed by molecular distillation from rapeseed oil
deodoriser distillate (RODD), a by-product of vegetable refining edible oil. The effects of evaporating
temperature, feed flow rate and wiper rolling speed on yield of tocopherol were systematically studied by
response surface methodology (RSM) based on the separation of fame previously. The first fraction, which
was collected at 50 1C at 2
66 Pa, contains mainly hydrocarbons, ketones and aldehydes. The second fraction,
which was collected at 100–110 1C at 5
32 Pa, contains mainly FAME and the content of FAME in the second
distillate was above 90% by analysis of gas chromatography–mass spectrometry (GC–MS). The fraction
collected at 200–230 1C at 2
66 Pa was rich in tocopherol (nearly 35%). The overall recovery of tocopherol and
FAME were around 50% and 90%, respectively, of the original content in RODD.
r 2006 Silsoe Research Institute. All rights reserved
Published by Elsevier Ltd

1. Introduction are palm oil for obtaining provitamin A, oils of rice for
the oryzanol recovery and natural tocopherols from
Molecular distillation is characterised by short vegetable-oil deodoriser distillate, a by-product of the
exposure of the distilled liquid to elevated temperatures, vegetable oil-refining process, composed of fatty acids,
high vacuum in the distillation space, and a small sterols, tocopherol, sterol esters, hydrocarbons, break-
distance between the condenser and evaporator (Shima- down products of fatty acids, aldehydes, ketones and
da et al., 2000; Ibanez et al., 2002). The short residence acyl glycerol species. Compounds such as tocopherol,
of the liquid on the evaporating cylinder, in the order of sterol and fatty acid methyl esters (FAME) are of
a few seconds to 1 min, is guaranteed by distributing the interest as high value-added products because of their
liquid in the form of a uniform thin film. By reducing the nutraceutical activities (Kusdiana & Saka, 2004).
pressure of non-condensable gas in the evaporator to Fatty acids constitute 25–75% of the distillate
lower than 0
1 Pa, a reduction in distillation tempera- depending on the raw material being refined, the type
tures can be obtained. Molecular distillation shows of refining process and the conditions employed therein.
promise in the separation, purification and concentra- Recently, fatty acids from deodoriser distillate are only
tion of natural products, usually composed of complex limited to the use of non-food, low-cost applications
and thermally sensitive molecules. Furthermore, this because they are contaminated. Within the past three
process has advantages over other techniques that use decades, interest in reducing air pollution and in
solvents as the separating agent, avoiding problems with developing domestic energy sources has encouraged
toxicity (Lutisan et al., 2002). research in many countries toward developing non-
The substitution of conventional materials used in the petroleum fuels for internal combustion engines; it has
nutrition, pharmaceutical, and cosmetic areas by natural been shown that the simple alcohol esters of fatty acids
products has gained interest and importance. Examples (FA) are acceptable alternatives diesel fuels. Biodiesel is

1537-5110/$32.00 383 r 2006 Silsoe Research Institute. All rights reserved

Published by Elsevier Ltd

转载
 ARTICLE IN PRESS
中国科技论文在线 http://www.paper.edu.cn
384 S.T. JIANG ET AL.

made from agricultural materials, which are produced standard. All other chemicals and reagents were of
via photosynthetic carbon fixation; its combustion does analytical grade. Rapeseed oil deodoriser distillate came
not contribute to net atmospheric carbon dioxide levels. from Fenda Oil and Fat Ltd (Hefei, China).
Its immediate application would be in the blend with
diesel, with a value that can reach up to 20%. Initial
efforts at producing biodiesel employed refined edible 2.2. Esterification of rapeseed oil deodoriser distillate
oil and animal fats as feedstocks for fuel synthesis. More
recently, methods have been developed to produce Rapeseed oil deodoriser distillate contained free fatty
FAME from cheaper material, such as spent restaurant acids (FFA), neutral glycerides including monoacylgly-
grease and waste edible oil (Hass et al., 2003). In this ceride (MAG), diacylglyceride (DAG) and triacylglycer-
study, the potentiality of the process for recovering ide (TAG), and unsaponifiable matters. Sulphuric acid
biodiesel (FAME) from rapeseed oil deodoriser distillate catalysed methylation of the FFA in deodoriser distillate
(RODD) was presented by converting the fatty acids to (methanol/deodoriser distillate in ml g1 at 100/100) was
biodiesel (FAME). conducted in vigorously shaken glass screw-capped
Some technical information is available pertaining to containers at 65 1C in the presence of sulphuric acid
the recovery of tocopherol, sterols and FAME, invol- (sulphuric acid/deodoriser distillate in ml g1 at 4/100)
ving a series of chemical and physical processing steps with N2 in gas (Shao et al., 2004).
(Ibanez et al., 2002; Mendes et al., 2002). These
techniques are used alone or in combination with 2.3. Crystallisation of sterols
others, including solvent extraction (Iglesias & Lozano,
2004), urea complexation (Udaya & Fereidoon, 1999). The yield and recovery for crystallisation of sterols
However, they involved energy consuming, labour- have been studied as follows: the acetone was used as
intensive and low recovery of tocopherol. Recently, it solvent, with crystallisation temperature at 5 1C and
was reported that lipase-catalysed methyl esterification crystallisation time of 3 h. A recovery of sterols as high as
followed by molecular distillation are very effective for 85% was achieved by using a single-stage batch crystal-
enrichment or purification of useful components in lisation of the solvent/material mixture (mass ratio of
deodoriser distillate (Ghosh & Bhattacharyya, 1996; 10:1) (Moreira et al., 2004).The products were defined as
Kusdiana & Saka, 2004; Ramamurthi & McCurdy. RODD fatty acids and tocopherol concentrate.
1993). However, the large scale of the lipase reaction is
difficult because of lipase characteristics and high costs
of lipase. Hence, industrial purification of tocopherols 2.4. Distillation of tocopherol and ester fraction
often adopts the acid-catalysed esterification to enhance
the profit in the developing countries. Few studies on the The FAME and tocopherol, obtained from RODD
optimisation of recovering tocopherol and FAME from sample by the combination of methyl esterification and
RODD have appeared in the literature. There appears crystallisation of sterols were fractionally distilled in a
to be a need for the optimisation of methyl esterification wiped-film molecular distillation (Fig. 1). The first
that is more suitable for the guidance of production. fraction, which was collected at 50 1C at 2
66 Pa,
Taking all of these concerns into consideration, in the contains mainly hydrocarbons, ketones and aldehydes.
present work, a molecular distillation process for The fraction from 100 to 140 1C at 5
32 Pa was collected
recovery of tocopherol and FAME from RODD has as the second fraction, which was rich in FAME. And
been put forward by a combination of acid-catalysed the third fraction was collected at 170–230 1C at 2
66 Pa
methyl esterification and crystallisation, followed by was rich in tocopherol and collected by changing the
fractional distillation of the derived product. collecting flask. The evaporating temperature, wiper
rolling speed, feed temperature and feed flow rate were
investigated systematically on the recovery of FAME
2. Materials and methods and tocopherol. The feed, residue and two fractions
were cooled to room temperature, weighed and analysed
2.1. Materials and reagents for FAME, tocopherol, and sterol content according to
the methods described below.
Palmitic, stearic, oleic, linoleic acid methyl esters,
a,g,d-tocopherol, b-sitosterol, stigmasterol, and campes- 2.4.1. Experimental design for the separation of fatty
terol were obtained from Sigma Chemical Co.(St Louis, acid methyl ester and tocopherol
USA). They were mixed in amounts proportional to The effects of evaporating temperature, feed flow rate,
their mass abundance in RODD which served as vacuum and wiper rolling speed were investigated on the
 ARTICLE IN PRESS
中国科技论文在线 http://www.paper.edu.cn
RECOVERING TOCOPHEROL AND FATTY ACID ESTERS 385

Table 1
Independent variables and their levels for central composite
b
a design

Independent variables Variable levels

1 0 +1

Evaporating temperature X1, 1C 170 200 230

Wiper rolling speed X2, min1 50 150 250
c Feed flow rate X3, ml min1 30 90 150

responses, namely Y1 and Y2. The model proposed for

d each response of Y was
h j Y ¼ A0 þ A1 X 1 þ A2 X 2 þ A3 X 3 þ A4 X 1 X 2 þ A5 X 1 X 3
e i
f
k þ A6 X 2 X 3 þ A7 X 21 þ A8 X 22 þ A9 X 23 ð2Þ
g
Fig. 1. Schematic diagram of wiped-film molecular still:(a) where: A0 is a constant; A1, A2, and A3 are linear
motor drive; (b) feed flask; (c) wiper; (d) residue receiver; (e) coefficients; A4, A5, and A6 are cross-product coeffi-
distillate receiver; (f) one-stage condenser inlet; (g) one-stage
condenser outlet; (h) two-stage condenser inlet; (i) two-stage
cients; and A7, A8 and A9 are quadratic coefficients.
condenser outlet; (j) cold trap; (k) rotary vane vacuum pump 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 obtained by holding one of the independent
separation of FAME by single factor analysis. A central variables at a constant value and changing the level of
composite design was employed to study the response, the other variables.
namely tocopherol concentration after distillation Y1 in
% by wt and recovery of tocopherol Y2 in % by wt [see
Eqn (1)]. The independent variables were X1, X2 and X3 2.5. Determination of acid value and saponification value
representing evaporating temperature, wiper rolling
speed and feed flow rate, respectively. The settings for Acid value and saponification value were determined
the independent variables were as follows (low and high by standard method (Ru, 2003).
values): evaporating temperature of 170 and 230 1C;
wiper rolling speed of 50 and 250 min1; feed flow rate
of 30 and 150 ml h1. Each variable to be optimised was 2.6. Gas chromatography—mass spectrometry analysis
coded at three levels: 1, 0 and +1. Three replicates at of fatty acid methyl ester
the centre (0,0,0) of the design were performed to allow
the estimation of the pure error. The central composite The FAME composition of each deodoriser distillate
design is shown in Table 1. All experiments were carried was determined by the following procedures. A Hewlett-
out in a randomised order to minimise the effect of Packard model HP 6890 gas chromatograph with a mass
unexpected variability in the observed response due to spectrometer (GC–MS), detector model (electron impact,
extraneous factors. 70 eV) and a DB-5 (30 m by 0
25 mm by 0
25 mm) capillary
column was used. Injection was carried out at 250 1C.
Y 2 ¼ ðT DR =T RO Þ  100 (1) Helium was the carrier gas. The oven temperature was
programmed from 180 1C for 3 min, and then from 180 to
where TDR and TRO are defined as the tocopherol 210 1C at 25 1C min1, from 210 to 280 1C at 25 1C min1.
content of tocopherol concentrate and RODD fatty The final temperature was maintained for 3 min. Some
acids and tocopherol concentrate, respectively. 2 ml of extracts obtained by wiped-film molecular distilla-
tion and diluted with hexane were injected. The target
2.4.2. Statistical analysis compounds were identified by mass spectrometric in both
As for the optimisation for purification of tocopherol SCAN (total number of ions) and selected ion monitoring
by molecular distillation, the responses were analysed (SIM) mode. Spectra of the compounds were obtained and
using Matlab software. A quadratic polynomial regres- compared with those in the US. National Institute of
sion model was assumed for predicting both Y Standards and Technology (NIST) library.
 ARTICLE IN PRESS
中国科技论文在线 http://www.paper.edu.cn
386 S.T. JIANG ET AL.

2.7. The gas chromatography analysis of fatty acid Table 2

methyl ester Composition of rapeseed oil deodoriser distillate

Component Content, %
The FAME content was determined with a gas
chromatography (Agilent 1790, Agilent, Shanghai, Sterol 14
14
Campesterol 3
52
China) connected to a DB-wax capillary column (30 m
b-Sitosterol 6
42
by 0
25 mm by 0
25 mm) using a tricaproin as an internal Stigmasterol 4
20
standard. The column temperature was controlled as Tocopherol 4
14
follows: 0
5 min at 190 1C, 4 min to 210 1C, 10 min to a-Tocopherol 0
98
280 1C, and 2 min at 280 1C. The injector and detector g-Tocopherol 2
05
d-Tocopherol 1
11
temperatures were 245 and 350 1C, respectively.

2.8. Analysis of tocopherol and sterol

of evaporating temperature on the recovery of FAME
and loss of tocopherol. The distilled FAME content
A Waters series high performance liquid chromoto-
increased with increasing temperature and reached
graphy (HPLC) equipped with photodiode array detec-
nearly steady state (39/100 g ml1) when the temperature
tor was used to analyse sterol and tocopherol in the
was 110 1C at 5
32 Pa. The tocopherol content in the
RODD and distillate obtained using the molecular
residue did not change much over the temperature range
distillation technique described above. The separation
tested. The effects of feed temperature were investigated
column was Atlantis (Waters Co., Ireland) of 15 cm by
next [Fig. 2(b)]. The temperature of a liquid entering the
0
25 mm by 0
25 mm. Samples were injected into the 20 ml
evaporating cylinder of a molecular evaporator was one
injection loop of the HPLC system. The mobile phase
of the important technological parameters that deter-
was methanol and water (96:4, ml ml1) at a flow rate of
mined evaporator operation and it was useful to gently
1 ml min1. Identification of compounds was achieved
pre-heat the feed before it entered the evaporator in
by comparing their retention timesd spectra with those
appropriate front-end heat exchanger to a temperature
of standards. For quantitative analysis, calibration
close to the asymptotic temperature. Figure 2(b) shows a
curves were prepared by analysing different concentra-
typical feed temperature course under these conditions.
tions of tocopherol and sterol standards and by
Recovery of FAME in the distillate increased to 0
385 g
representing peak area versus concentration. Detection
ml1. When the feed temperature was 80 1C, the loss of
for tocopherol and sterols was performed at their
tocopherol was about 0
12%. The effects of rolling
adsorption maxima in the ultraviolet, that is, 296 and
speed of wiped-film molecular still were investigated in
205 nm, respectively (Abidi & Mounts, 1997).
Fig. 2(c). For the rolling speed used, it can be seen that
Tocopherol was expressed as a total tocopherol
when the rolling speed is 150 min1, the loss of
content (sum of a,g, and d-tocopherol) and the sterols
tocopherol was very small although FAME could not
as a sum of total sterol. The amounts of hydrocarbons
eliminated. Based on the analysis above, the reaction
and other components were estimated by different of
conditions were determined as follows: evaporating
total sterols and tocopherol from unsaponifiable matter.
temperature 110 1C at 5
32 Pa, rolling speed of
150 min1 and feed temperature of 80 1C. The esters
can be practically removed from the feed stream,
3. Results and discussion
thereby concentrating the tocopherol by molecular
distillation. The generated biodiesel presented light
3.1. Composition of rapeseed oil deodoriser distillate
colour, similar to the refined rapeseed oil, and the
colour number was determined according to GB252
94
The results of sterols, tocopherols in RODD before
of China. Here, the focus was on the tocopherol
distillation are shown in Table 2. The acid value and
recovery, this one much more thermally sensitive than
saponification value of RODD were 80
70 and
biodiesel. The biodiesel recovery was easier and simpler.
155
98 mg [KOH] g1, respectively.
Total ion chromatography of FAME was illustrated in
Fig. 3 by analysis of GC–MS. The components of
3.2. The recovery of fatty acid methyl ester from rapeseed biodiesel from deodoriser distillate and part properties
oil deodoriser distillate were shown in Tables 3 and 4, respectively. The content
of biodiesel accorded with the rapeseed oil components.
Several factors affecting the FAME treatment of Palmitic methyl ester was the main fatty acid methyl
RODD are shown in Fig. 2. Figure 2(a) shows the effect ester (33
61%) and erucic methyl ester (7
39%) was
 ARTICLE IN PRESS
中国科技论文在线 http://www.paper.edu.cn
RECOVERING TOCOPHEROL AND FATTY ACID ESTERS 387

48 4 3.3.1. Model fitting

Recovery of FAME,

Table 5 showed the responses, i.e. the concentration

Tocopherol content
in the distillate, %
40 3 of tocopherol Y1 and recovery of tocopherol Y2 after
g 100 ml −1

molecular distillation. The response and variables were

32 2
fitted to each other by multiple regressions. A good fit
24 1 was obtained.
Y 1 ¼ ð35 þ 0
12125X 1 þ 0
14125X 2  0
2875X 3
16 0
90 100 110 120 130 140  0
2425X 1 X 2 þ 0
35X 1 X 3  0
025X 2 X 3
Evaporating temperature of  2
94125X 21  0
56625X 22  1
25875X 23 Þ=100
(a) wiped-film molecular still, °C
ð3Þ
40 Y 2 ¼ ð87
833 þ 6
2X 1  1
875X 2  0
5X 3
0.20
Recovery of FAME,

Tocopherol content
in the distillate, %
 0
05X 1 X 2 þ 0
45X 1 X 3 þ 1
05X 2 X 3
g 100 ml −1

38 0.16  4
266X 21  2
266X 22  6
01667X 33 Þ=100

ð4Þ
36 0.12
The value for R2 and small probability P (Po0
05
0.08 when significant) were 0
9176 and 0
0294 for Eqn (3),
34
40 50 60 70 80 90 and 0
9378 and 0
01533 for Eqn (4), respectively. Also,
Feed temperature of the predicted results, according to models for concen-
(b) wiped-film molecular distillation, °C tration and recovery of tocopherol, were close to the
observed experimental responses. These indicated that
42 1.8 the generated models adequately explained the data
Tocopherol content
Recovery of FAME,

in the distillate, %

variation and significantly represented the actual re-

41
1.5 lationships between the reaction parameters.
g 100 ml −1

40
1.2 3.3.2. Effects of parameters
39 Many parameters can influence the separation per-
0.9 formance of tocopherol. The most important para-
38 meters were evaporator temperature, feed flow rate,
40 80 120 160 200 240
vacuum and wiper rolling speed. For practical opera-
Rolling speed of wiped-film
tion, the lowest vacuum (2
66 Pa) should be used to
(c) molecular still, min−1
maximise the separation efficiency.
Fig. 2. The effects on the recovery of biodiesel (m) and Equations (3) and (4) showed that concentration and
tocopherol (J) of: (a) evaporating temperature; (b) feed recovery of tocopherol have a complex relationship with
temperature and (c) wiper rolling speed independent variables that encompass both first- and
second-order polynomials and may have more than one
lower, because the raw material was the low erucic acid maximum point. Analysing the contour plots for
rapeseed seed. From Table 4, it was shown the obtained concentration and recovery of tocopherol was the best
FAME in the present study was acceptable as an way to evaluate the relationships between responses,
alternative diesel fuel. variables and interactions that existed herein. The three-
dimensional response surfaces were plotted in Figs 4–7,
as a function of the interactions of any two of the
3.3. The recovery of tocopherol from rapeseed oil variables by holding the other one at high value. All four
deodoriser distillate plots in Figs 4–7 showed similar relationships with
respect to the effects of each variable. The response
Since FAME were the most volatile components and obtained were convex nature suggesting that there were
were recovered first at lower temperature, they were well-defined optimum operating conditions. However,
omitted in order to facilitate the recovery of tocopherol. the convexity was not high enough, as the surfaces were
By increasing the evaporating temperature, it was rather symmetrical and a little flat near the optimum
possible to recover in the vapour phase with MAG which meant that the response optimised value based on
and DAG. Phytosterols were recovered in the liquid combined effects evaporating temperature, feed flow
phase also called residue. rate and feed temperature may not vary widely from the
 ARTICLE IN PRESS
中国科技论文在线 http://www.paper.edu.cn
388 S.T. JIANG ET AL.

single variable optimised conditions. Figures 4 and 6 whereas, in Figs 4, 5 and 7, evaporating temperature
revealed that wiper rolling speed did not show sig- showed significant variation both above and below the
nificant variation on the concentration of tocopherol, optimum values. With the increase of temperature, an

8.65
100 1295

11.28
336
Relative intensity, %

10.97
799
14.32
298

12.83
126
7.18
137

3.64 5.54
11 12

0
5.00 10.00 15.00 20.00 25.00
Time, min
Fig. 3. Total ion gas chromatogram of fatty acid methyl ester

Table 3
The components of biodiesel from deodoriser distillate and relative content

t, min Component Molecular formula Molecular weight Relative content, %

7
18 Hydrocarbon C20H38 278
2990 4
01

8
65 Palmitic methyl ester C17H34O2 270
2559 33
61
10
97 Linoleic methyl ester C19H34O2 294
2559 18
25
11
08 Oleic methyl ester C19H36O2 296
2694 16
14
11
28 Stearic methyl ester C19H38O2 298
2872 8
81
12
83 Gondoic methyl ester C21H40O2 324
3005 3
78
14
32 Erucic methyl ester C23H44O2 352
3341 7
39

Note: The content of erucic acid methyl ester in rapeseed oil deodoriser distillate is 3
02%.

Table 4
The properties of biodisel, No.0 diesel and their mixture

Physical–chemical properties Biodiesel No.0 diesel(GB252
94) Mixture of biodiesel and diesel

Density at 25 1C, kg m3 878
25 858
32 860
75

Viscosity at 25 1C, mPa s2 8
00 7
36 7
54
Cold filter plugging point, 1C 4
00 4
00 2
00
Colour number 15
00 16
00 16
00
Acid value, mg [KOH] g1 0
66 0
49 0
54

Note: Colour number means the colour of diesel determined by GB252
94, China. The mixture of biodiesel and diesel is composed of biodiesel
and No.0 diesel with ratio of 1:1 (v/v).
 ARTICLE IN PRESS
中国科技论文在线 http://www.paper.edu.cn
RECOVERING TOCOPHEROL AND FATTY ACID ESTERS 389

Table 5
Central composite design and experiment data

Run Independent variables Responses

Evaporating temperature Wiper rolling speed Feed flow rate (X3), Y1 % Y2 %

(X1), 1C (X2), min1 ml min1

1 0 1 1 33
50 85
0
2 0 1 1 34
63 82
0
3 0 1 1 33
20 75
0
4 0 1 1 32
80 76
2
5 1 0 1 30
90 72
0
6 1 0 1 29
40 70
0
7 1 0 1 31
50 84
2
8 1 0 1 31
40 84
0
9 1 1 0 31
20 75
2
10 1 1 0 32
60 75
7
11 1 1 0 30
87 87
0
12 1 1 0 31
30 87
3
13 0 0 0 35
40 87
4
14 0 0 0 34
70 88
1
15 0 0 0 34
90 88
0

250 33
32
31 .5

31. 5
32.5

32

32
30

33
30.5

33

.5
31

31

32
140

31
31
Rolling speed, min−1

200

32
33
Feed flow rate, ml h−1

33
120 34
33

34
31.5
32.5
32
31.5

32
33
32.5

100
32

150
30.5

31
31

34

80

32
33
33

100
33

60

34
31 .5
32.5
32

33
31. 5

32
32
29. 5

.5

32
30. 5

31
31

31
34
30

50 40
170 180 190 200 210 220 230

32
33
170 180 190 200 210 220 230
Evaporating temperature, °C
Evaporating temperature, °C

33 34.5
34 34
32.5 35
33.5
Concentration of

33
tocopherol, %

32 34
Concentration of
tocopherol, %

32 33
33
31.5 32.5
31 32
31 31 32
30
30.5 30 31.5
29
250 29 31
30 150
Ro 200 240 240
llin Fee 30.5
gs 150 200
220 29.5 df 100
low
220
pe 100 50 200 30
ed 180 ting rate 180 ting
,m 50 160 pora °C ,m pora
Eva ature, °
160 C
in −1 Eva rature, l h −1 0
p e p e r
tem tem

Fig. 4. Contour plot and response surface curve showing Fig. 5. Contour plot and response surface curve showing
predicted response surface of yield of tocopherol as a function predicted response surface of yield of tocopherol as a function
of evaporating temperature and rolling speed (feed flow rate of of evaporating temperature and feed flow rate (wiper rolling rate
150 ml h1, feed temperature of 80 1C) of 250 min1, feed temperature of 80 1C)
 ARTICLE IN PRESS
中国科技论文在线 http://www.paper.edu.cn
390 S.T. JIANG ET AL.

30 8
. 31. 31 30 .8 30 .6 250 80

68
70
31 2 31 .2
140

74
31.4

76

78
31. 31 .4
31. 2 31.6 31 82

72
4

Wiper rolling speed, min−1

31 .
31 1. 6 31.8 31 .8 6
Feed flow rate, ml h−1

120 3 31 . 200 82
32 4

80
8
31 .

100 32

31.6

78
31.8

76
74
32

31 .8
150
32

80

72

82
32 .4
31 . 31 .6 31.8 31
60 4 31. 8 .
31 6

80
31 . 31.6 100
31 2 31.4 31..4 31
31 2

78
40 31.2

76
74
70
.
30 .6 30 .8 31 31 30 8 30. 6
82

72
50 100 150 200 250 50
Wiper rolling speed, min−1 170 180 190 200 210 220 230
Evaporating temperature, °C

32
31.8

Recovery of tocopherol, %
32.5 82
85
31.6
Concentration of

32 80
tocopherol, %

31.5 31.4 80
78
31.2 75
31 76
31
30.5 70
30.8 74
30
150
250
30.6 65 72
Fee 100 250
df 200 30.4 200 240 70
low 50 150 220
rat 100 lling Wi
p 150 200
e, m 0 50 p er ro in−1 spe er rol 100 ing 68
i orat , °C
l h −1 W d, m 180
ed, ling 50 160 a p
spee min − Ev rature
1 pe
tem
Fig. 6. Contour plot and response surface curve showing
predicted response surface of yield of tocopherol as a function Fig. 7. Contour plot and response surface curve showing
of wiper rolling speed and feed flow rate (evaporating predicted response surface of recovery of tocopherols as a
temperature of 230 1C, feed temperature of 80 1C) function of evaporating temperature and wiper rolling speed
(feed flow rate of 150 ml h1, feed temperature of 80 1C)

increase in the amount of condensate in the cold trap distillate). When the operating temperature was in-
was observed; therefore, this may influence the risk of creased to 230 1C, part of phytosterol and MAG began
evaporator temperature of volatiles into the pumps. A to be recovered in the vapour phase too; however, this
temperature about 200 1C was therefore suggested. was not of interest because at this temperature, the
The effect of wiper rolling speed was tested. The phytosterols started to distillate and this reduces the
function of the roller was to distillate the material evenly purity of tocopherol.
on the surface of the internal heating wall and to control
the thick of the material film on the wall. There a certain 3.3.3. Verification of the model
speed should be used to efficiently supply the materials By partial differential coefficient operation to this
and maintain surface because noise was brought out model, and equal to zero, then the point of curved
when fast wiper rolling speed is applied. From Figs 4 surface was determined (X 1 ¼ 0; X 2 ¼ 0; X 3 ¼ 0; Y 1 ¼
and 6, when the wiper rolling speed was about 35%; Y 2 ¼ 87
833%). According to the response sur-
150 min1, tocopherol content was above 31%. face methodology (RSM) result, an experiment with an
Recovery of tocopherols as a function of evaporating evaporating temperature of 200 1C, feed flow rate of
temperature and wiper rolling speed (feed flow rate of 90 ml h1 and wiper rolling speed of 150 min1 was
90 ml h1) is shown in Fig. 7. It can be observed in Fig. 4 conducted in order to investigate the effect of RSM. The
that for a feed flow rate of 90 ml h1 and evaporator experiment was carried out at the optimised conditions.
temperature of 200 1C, the tocopherol content was The content and recovery of tocopherol was above
almost 35% recovered in the vapour phase (also called 34
7%, 89
5% after molecular distillation and was in
 ARTICLE IN PRESS
中国科技论文在线 http://www.paper.edu.cn
RECOVERING TOCOPHEROL AND FATTY ACID ESTERS 391

good agreement with the predicted one. The model Hass M J; Michalski P J; Runyon S; Nunez A; Scott K M
equation developed can be used for predicting purifica- (2003). Production of FAME from acid oil, a by-product of
tion of tocopherol. vegetable oil refining. Journal of the American Oil Chemist’s
Society, 80, 97–102
Ibanez E; Benavides A M H; Senorans F J; Reglero G (2002).
Concentration of sterols and tocopherols from olive oil with
4. Conclusions supercritical carbon dioxide. Journal of the American Oil
Chemist’s Society, 79, 1255–1260
From the data obtained in this study, a process was Iglesias M T; Lozano J E (2004). Extraction and characteriza-
proposed for recovering fatty acid methyl esters tion of sunflower pectin. Journal of Food Engineering, 62,
(FAME) (purity: 490%) and tocopherol (pur- 215–223
ity:435%) from rapeseed oil deodoriser distillate. The Kusdiana D; Saka S (2004). Two-step preparation for catalyst-
overall recovery of tocopherols and FAME were around free biodiesel fuel production. Applied Biochemistry and
Biotechnology, 113, 781–790
50% and 90%, respectively, of the original content in
Lutisan J; Cvengros J; Micov M (2002). Heat and mass transfer
raw material. Model of separation of tocopherol was in the evaporating film of a molecular evaporator. Chemical
developed on the basis of the analysis of response Engineering Journal, 85, 225–234
surface methodology. The evaporating temperature was Mendes M F; Pessoa F L P; Uller A M C (2002). An economic
the most important parameter about recovery of FAME evaluation based on an experimental study of the vitamin E
and tocopherol. This process was feasible from an concentration present in deodorizer distillate of soybean oil
economic point of view and was a promising measure using supercritical CO2. Journal of Supercritical Fluids, 23,
for further utilisation of agricultural products. 257–265
Moreira E A; Baltanas M A (2004). Recovery of phytosterols
from sunflower oil deodorizer distillates. Journal of the
American Oil Chemist’s Society, 81, 161–167
Acknowledgements Ramamurthi S; McCurdy A R (1993). Enzymatic pretreatment
of deodorizer distillate for concentration of sterols and
This work was supported by Natural Science Foun- tocopherols. Journal of the American Oil Chemist’s Society,
dation of Anhui Province (no. 03041302). And the 70, 287–295
authors gratefully acknowledge the contribution of the Ru Y J (2003). Analysis of Oil and Fat. Chinese Chemistry
following to this work: Y. Li and X.Y. Zhang for their Industry Press, Beijing, China
value advice and Associate Professor Y.K. Ye for his Shao P; Jiang S T; Dai Y (2004). Application of the response
surface method based on neutral networks to methanol
comments on this manuscript.
esterification of rapeseed deodorizer distillate. Chinese Food
Science (Suppl.), 25, 78–81
References Shimada Y; Nakai S; Suenaga M; Sugihara A; Kitano M;
Tominaga Y (2000). Facile purification of tocopherols from
Abidi S L; Mounts T L (1997). Reversed-phase high- soybean oil deodorizer distillate in high yield using lipase.
performance liquid chromatographic separations of toco- Journal of the American Oil Chemist’s Society, 77,
pherols. Journal of Chromatography A, 782, 25–32 1009–1013
Ghosh S; Bhattacharyya D K (1996). Isolation of tocopherol Udaya N W; Fereidoon S (1999). Concentration of 3-
and sterol concentrate from sunflower oil deodorizer polyunsaturated fatty acids of seal blubber oil by urea
distillate. Journal of the American Oil Chemist’s Society, complexation: optimization of reaction conditions. Food
73, 1271–1274 Chemistry, 65, 41–49