www.elsevier.com/locate/jfoodeng
a
Department of Food Engineering, University of Gaziantep, Gaziantep TR-27310, Turkey
Department of Food Science and Technology, One Shields Avenue, University of California Davis, Davis, CA 95616, United States
Received 15 March 2007; received in revised form 14 May 2007; accepted 15 May 2007
Available online 18 May 2007
Abstract
Blends of barley our and tomato pomace were processed in a co-rotating twin-screw extruder. Experimental design with die temperature (140160 C), screw speed (150200 rpm) and tomato pomace level (210%) as independent variables produced 20 dierent combinations that were studied using response surface methodology to investigate the eect of these variables on system parameters (SME,
die melt temperature and die pressure) and product responses (expansion, bulk density, water absorption and solubility indices, texture
and color). Extrudate from ve experiments within 20 samples was selected for sensory evaluation in terms of color, texture, taste, oodor and overall acceptability. Regression equations describing the eect of each variable on the system parameters and product
responses were obtained. The system parameters and product responses were most aected by changes in temperature, pomace level
and to a lesser extent by screw speed. Extrudates with 2% and 10% tomato pomace levels extruded at 160 C and 200 rpm had higher
preference levels for parameters of color, texture, taste and overall acceptability. The results suggest that tomato pomace can be extruded
with barley our into an acceptable and nutritional snack.
2007 Elsevier Ltd. All rights reserved.
Keywords: Extrusion cooking; Barley; Tomato pomace; Response surface methodology
1. Introduction
Tomato (Lycopersicon esculentum) is one of the most
popular vegetables and an integral part of human diet
worldwide. Signicant amounts are consumed in the form
of processed products such as juice, paste, puree, ketchup,
sauce and salsa. During tomato processing a by-product,
known as tomato pomace, is generated. This by-product
represents, at most, 4% of the fruit weight (Del Valle,
Camara, & Torija, 2006). Tomato pomace consists of the
dried and crushed skins and seeds of the fruit (Tadeu-Pontes, Carvalheiro, Roseiro, & Amaral-Colloco, 1996). The
skin, important component of pomace, is source of lycopene. Lycopene is an excellent natural food color and also
Corresponding author. Tel.: +90 342 3172309; fax: +90 342 3601105.
E-mail address: maskan@gantep.edu.tr (M. Maskan).
0260-8774/$ - see front matter 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jfoodeng.2007.05.014
serves as a functional ingredient with important health benets beyond basic nutrition (Kaur, Sogi, Gary, & Bawa,
2005). A diet rich in lycopene is related to a decreased risk
of certain cancers, particularly cancers of the digestive
tract, prostate cancer and pancreatic cancer due to protective eect of lycopene against oxidative damage (Johnson,
2000). It also was found that tomato pomace signicantly
reduced cholesterol level in liver and heart by 15% and
18%, respectively (Bobek, Ozdin, & Hromadova, 1998).
The use of tomato processing by-products could provide
gaining valuable substances and at the same time reduce
the waste disposal problem.
Dietary ber has received increased attention recently.
As consumers become more concerned about eating food
with health benets, barley, which is naturally healthy, easily available and inexpensive crop is strongly favored for
increased incorporation into human diet (Czuchajowska,
Klamczynski, Paszczynska, & Baik, 1998). The dietary
232
233
Table 1
Process variables used in the central composite design for three independent variables
Code
X1
X2
X3
1
1.682
133.18
133
0
140
150
2
150
175
6
160
200
10
166.82
217
12.73
Table 2
Experimental design for extrusion experiment with coded and actual
variable levels
Run
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Coded levels
Actual levels
X1
X2
X3
Die
temperature
(C)
Screw
speed
(rpm)
Pomace
level (%)
1
1
1
1
1
1
1
1
1.682
1.682
1
1
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
1.682
1.682
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
0
0
1.682
1.682
0
0
0
0
0
0
140
160
140
160
140
160
140
160
133.18
166.82
150
150
150
150
150
150
150
160
150
150
150
150
200
200
150
150
200
200
175
175
133
217
175
175
175
175
175
175
175
175
2
2
2
2
10
10
10
10
6
6
6
6
0
12.73
6
6
6
6
6
6
234
S e W e he
;
S d W d hd
where Se and Sd are the cross-sectional areas of the extrudate and the die; We and he are the width and thickness of
the extrudate and Wd and hd are the width and thickness of
the die, respectively.
2.6.2. Bulk density
Bulk density was determined by measuring the volume
of extrudate by glass bead displacement (Hwang & Hayakawa, 1980; Sokhey, Ali, & Hanna, 1997). Glass beads with
a diameter range of 1.001.18 mm were used as displacement medium. Bulk densities of the extrudates were calculated as
qb
W ex
qgb ;
W gb
2.6.4. Texture
The hardness of samples was measured with a TA-XT2i
Texture Analyzer (Texture Technologies Corp., Scarsdale,
NY, USA). Hardness in N was determined by measuring
the maximum force required to break the extruded samples
(42 mm long) using three point bend test with a sharpbladed probe (55 mm wide, 40 mm high, 9 mm thick).
The test speed was 2 mm/s and the distance between two
supports was 22 mm. A forcetime curve was recorded
and analyzed by Texture Exponent 32 software program
(version 3.0). Elevan measurements were performed on
each sample and averaged.
2.6.5. Color
HunterLab LabScan XE (Hunter Associates Laboratory, Inc., Reston, VA, USA) was used to determine color
values of the raw materials and ground extruded in terms
of the L, a and b as measures of lightness, redness and yellowness, respectively. The measuring head was equipped
with 51 mm diameter viewing port and used the system
of diuse illumination with 10 viewing geometry. The
illuminant was D65. The colorimeter was calibrated
against a standard white tile (L = 91.43, a = 0.74,
b = 0.25). The extrudates were ground in a laboratory
grinder and passed through a 60 mesh sieve prior to color
analysis. For each sample, four measurements were taken
and averaged. The total color change (DE) was calculated
as
q
2
2
2
DE L L0 b b0 a a0 ;
4
where the subscript 0 indicates initial color values of the
raw material.
2.7. Sensory evaluation
A semi-trained panel of 34 students and faculty from
Food Engineering Department evaluated the extruded
snacks for color, texture and overall acceptability on a 7point hedonic scale (from 1 = extremely dislike to
7 = extremely like), while taste in terms of bran, tomato
avor and bitterness and o-odor was rated on a 7-point
scale (from 1 = none to 7 = very high). Panelists rinsed
their mouths with water after tasting each sample.
3. Results and discussion
Figures for die pressure, expansion, WAI, L, a and b
were not given for the sake of simplicity.
3.1. Diagnostic checking of tted model and surface plots for
various responses
3.1.1. Specic mechanical energy
A regression analysis were carried out to t mathematical models to the experimental data. The predicted model
Response
Source
df
Sum of
squares
Mean
squares
Fvalue
SME
Regression
Lack-of-t
Pure error
Residual
9
5
5
10
46169.46
610.87
3225.64
3836.51
5129.94
122.17
645.13
383.65
13.37
0.19
0.0002*
0.9541
Total
19
50005.97
Regression
Lack-of-t
Pure error
Residual
6
8
5
13
3.820 107
5.201 106
2.928 105
5.493 106
6.367 106
6.501 105
58551.11
4.226 105
15.07
11.10
<0.0001*
0.0084*
Total
19
4.370 107
P-value
SME (Wh/kg)
Table 3
Analysis of variance results for tted models
235
334.68
305.46
276.24
247.02
217.80
10.0
200.0
8.0
187.5
6.0
Pomace level (%) 4.0
2.0 150.0
175.0
162.5 Screw speed (rpm)
236
Table 4
Correlation coecients between product responses and system parameters
SEI
SEI
BD
WAI
WSI
L
a
b
DE
H
SME
P
T
BD
WAI
ns
0.219
1
WSI
ns
0.149
0.212ns
1
L
ns
0.149
0.542*
0.184ns
1
a
*
0.502
0.106ns
0.508*
0.462*
1
b
*
0.512
0.113ns
0.548*
0.433ns
0.988**
1
DE
ns
0.441
0.180ns
0.543*
0.476*
0.974**
0.992**
1
ns
0.236
0.376ns
0.269ns
0.309ns
0.721**
0.722**
0.752**
1
SME
ns
0.282
0.925**
0.397ns
0.467*
0.031ns
0.038ns
0.110ns
0.319ns
1
P
ns
0.382
0.152ns
0.133ns
0.782**
0.271ns
0.212ns
0.240ns
0.131ns
0.134ns
1
T
**
0.754
0.351ns
0.188ns
0.488*
0.089ns
0.080ns
0.171ns
0.235ns
0.444ns
0.564**
1
0.494*
0.644**
0.091ns
0.618**
0.296ns
0.296ns
0.364ns
0.323ns
0.637**
0.533*
0.777**
1
SEI: sectional expansion index; BD: bulk density; WAI: water absorption index; DE: total color change; H: hardness; SME: specic mechanical energy;
P: die pressure; T: die melt temperature.
ns
Not signicant.
*
Signicant at P < 0.05.
**
Signicant at P < 0.01.
The negative coecient of the rst order term of temperature (X1) (Eq. (6)) indicated that die pressure increased with
decrease of temperature. Meanwhile, negative coecient of
interaction term (X1X3) of temperature and pomace level
also resulted in decrease of die pressure. ANOVA for the
2FI model as tted to experimental results (Table 3) shows
signicance (P < 0.05). The coecient of determination
(R2) for die pressure was 0.8743. Die pressure model
showed signicant (P < 0.05) lack-of-t. The measured
die pressure in extrusion cooking of barley our and tomato pomace blends ranged from 786.45 to 6106.29 kPa. The
pressure at the die exit was decreased upon increase in temperature. Decrease in die pressure with the increase in temperature may be attributed to decrease in viscosity of the
melt (Ryu & Ng, 2001; Singh & Smith, 1997) due to degradation of gelatinized starch granules (Cai, Diosady, & Rubin, 1995; Singh, Sekhon, & Singh, 2007). A negative
correlation was found between die pressure and melt temperature (R = 0.777, P < 0.01) (Table 4). It was observed
that increasing pomace level with increasing temperature
decreased die pressure.
3.1.3. Expansion
The regression equation for expansion as sectional
expansion index (SEI) at any temperature (X1) and pomace
level (X3) was
SEI 1:59 0:25X 1 0:18X 3 0:14X 21 0:073X 23
0:099X 1 X 3 :
237
Table 5
Analysis of variance results for tted models
Response
Source
df
Sum of squares
Mean squares
F-value
P-value
SEI
Regression
Lack-of-t
Pure error
Residual
9
5
5
10
1.71
0.071
7.913 103
0.079
0.19
0.014
1.583 103
7.914 103
23.95
9.00
<0.0001*
0.0154*
Total
19
1.78
Regression
Lack-of-t
Pure error
Residual
9
5
5
10
0.50
0.033
7.699 103
0.041
0.055
6.611 103
1.540 103
4.075 103
13.58
4.29
0.0002*
0.0679
Total
19
0.54
Regression
Lack-of-t
Pure error
Residual
3
11
5
16
0.76
0.46
0.15
0.60
0.25
0.041
0.030
0.038
6.70
1.40
0.0039*
0.3741
Total
19
1.36
Regression
Lack-of-t
Pure error
Residual
3
11
5
16
44.94
8.47
0.67
9.14
14.98
0.77
0.13
0.57
26.23
5.74
<0.0001*
0.0333*
Total
19
54.08
Regression
Lack-of-t
Pure error
Residual
9
5
5
10
742.88
9.34
0.053
9.39
82.54
1.87
0.011
0.94
87.91
176.61
<0.0001*
<0.0001*
Total
19
752.27
BD
WAI
WSI
0.836
0.719
0.602
0.484
0.367
10.0
140.0
145.0
8.0
150.0
Temperature ( C) 155.0
4.0
6.0
Pomace level (%)
160.0 2.0
Fig. 3. Response surface plot for bulk density (BD) as a function of
pomace level and temperature at a screw speed of 175 rpm.
9
10
11
11.83
WSI (%)
238
10.74
9.66
8.57
7.48
200.0
160.0
187.5
155.0
175.0
150.0
145.0
Temperature (C)
150.0 140.0
Fig. 4. Response surface plot for water solubility index (WSI) as a
function of temperature and screw speed at a pomace level of 6%.
(X1) and screw speed (X2) at P < 0.001 and P < 0.05,
respectively. Temperature had also signicant quadratic effect (P < 0.001) on hardness of extrudates. The interaction
term (X1X3) between temperature and pomace level was
signicant, so that high values of hardness were found at
high level of pomace, dependent on temperature. Regression model (Eq. (11)) tted to experimental results of hardness showed higher coecient of determination
(R2 = 0.9875). Table 5 shows that the F-value for hardness
was signicant with a signicant lack-of-t (P < 0.05).
The textural property of barley our and tomato pomace extrudate was determined by measuring the force
required to break the extrudate (Singh, Hoseney, & Faubion, 1994). The higher the value of maximum peak force
required in gram, which means the more force required
to breakdown the sample, the higher the hardness of the
sample to fracture (Li, Zhang, Tony Jin, & Hsieh, 2005).
The eect of temperature and tomato pomace level on
hardness of extrudates is shown in Fig. 5. Response surface
plot showed that a decrease in die temperature with
increasing level of tomato pomace increased the product
hardness. Hardness of barley our and tomato pomace
extrudate varied between 5.64 and 29.75 N. A decrease in
die temperature increased the product hardness giving a
maximum at about 133.18 C, 175 rpm screw speed and
6% tomato pomace level. This result is in line with density
where an increase in density was observed. High density
product naturally oers high hardness evident by high correlation between product density and hardness (R = 0.925,
P < 0.01) (Table 4). Similar eect of temperature has been
observed while extruding yam and wheat our (Ding et al.,
2006; Sebio & Chang, 2000). Increasing screw speed
slightly decreased the hardness of the barley ourtomato
pomace extrudate, particularly at higher temperatures.
Liu, Hsieh, Heymann, and Hu (2000) found that the
hardness of the extruded oatcorn our increased as the
screw speed decreased.
Table 6
Analysis of variance results for tted models
Response
Source
df
Sum of
squares
Mean
squares
L-value
Regression
Lack-of-t
Pure error
Residual
9
5
5
10
139.17
1.98
2.21
4.19
15.46
0.40
0.44
0.42
36.91
0.90
<0.0001*
0.5468
Total
19
143.36
Regression
Lack-of-t
Pure error
Residual
9
5
5
10
138.73
0.42
0.28
0.71
15.41
0.085
0.056
0.071
218.42
1.50
<0.0001*
0.3335
Total
19
139.44
Regression
Lack-of-t
Pure error
Residual
9
5
5
10
113.23
0.68
0.13
0.82
12.58
0.14
0.027
0.082
154.03
5.07
<0.0001*
0.0497*
Total
19
114.05
Regression
Lack-of-t
Pure error
Residual
9
5
5
10
15.57
1.20
1.32
2.52
6.85
0.91
0.0029*
0.5409
Total
19
18.09
a-Value
b-Value
DE
Hardness (N)
16.33
12.47
8.60
4.74
140.0
10.0
145.0
8.0
6.0
150.0
Temperature (C) 155.0
Pomace level (%)
4.0
160.0 2.0
Fig. 5. Response surface plot for hardness (H) as a function of pomace
level and temperature at a screw speed of 175 rpm.
1.73
0.24
0.26
0.25
Fvalue
P-value
3.1.7. Color
Color is an important quality factor directly related to
the acceptability of food products, and is an important
physical property to report for extrudate products. The
regression equations for color parameters (L, a and b values) and total color change (DE) at any temperature (X1)
and pomace level (X3) were
L 69:95 3:23X 3 0:79X 23 ;
a 9:88 3:21X 3
20.20
239
12
0:92X 23 ;
13
0:86X 23 ;
14
15
240
9.61
E
9.04
8.48
7.91
7.34
Table 7
Sensory evaluation scores of extrudates produced at dierent conditions
Extrudates
AA
Color
Texture
3.64a
3.91a
4.70bc
5.32b
4.85bc
4.58c
5.08c
5.26b
4.23ab
3.44a
Taste
Bran avor
Tomato avor
Bitterness
3.67a
1.67a
1.85a
3.79a
1.50a
1.61a
3.23a
2.70b
1.94a
3.44a
2.50b
1.88a
3.82a
3.02b
2.00a
O-odor
Overall acceptability
1.79a
3.94a
1.79a
4.85b
1.70a
4.94b
1.58a
5.23b
1.88a
4.08a
140.0
10.0
145.0
8.0
150.0
6.0
Pomace level (%)
155.0
4.0
Temperature ( C)
160.0 2.0
Fig. 6. Response surface plot for total color change (DE) as a function of
pomace level and temperature at a screw speed of 175 rpm.
241
242
All in-text references underlined in blue are linked to publications on ResearchGate, letting you access and read them immediately.