Você está na página 1de 12

Journal of Food Engineering 84 (2008) 231–242


Evaluation of snack foods from barley–tomato pomace blends

by extrusion processing
Aylin Altan a, Kathryn L. McCarthy b, Medeni Maskan 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


Blends of barley flour and tomato pomace were processed in a co-rotating twin-screw extruder. Experimental design with die temper-
ature (140–160 °C), screw speed (150–200 rpm) and tomato pomace level (2–10%) as independent variables produced 20 different com-
binations that were studied using response surface methodology to investigate the effect 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 five experiments within 20 samples was selected for sensory evaluation in terms of color, texture, taste, off-
odor and overall acceptability. Regression equations describing the effect of each variable on the system parameters and product
responses were obtained. The system parameters and product responses were most affected 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 flour into an acceptable and nutritional snack.
Ó 2007 Elsevier Ltd. All rights reserved.

Keywords: Extrusion cooking; Barley; Tomato pomace; Response surface methodology

1. Introduction serves as a functional ingredient with important health ben-

efits beyond basic nutrition (Kaur, Sogi, Gary, & Bawa,
Tomato (Lycopersicon esculentum) is one of the most 2005). A diet rich in lycopene is related to a decreased risk
popular vegetables and an integral part of human diet of certain cancers, particularly cancers of the digestive
worldwide. Significant amounts are consumed in the form tract, prostate cancer and pancreatic cancer due to protec-
of processed products such as juice, paste, puree, ketchup, tive effect of lycopene against oxidative damage (Johnson,
sauce and salsa. During tomato processing a by-product, 2000). It also was found that tomato pomace significantly
known as tomato pomace, is generated. This by-product reduced cholesterol level in liver and heart by 15% and
represents, at most, 4% of the fruit weight (Del Valle, 18%, respectively (Bobek, Ozdin, & Hromadova, 1998).
Camara, & Torija, 2006). Tomato pomace consists of the The use of tomato processing by-products could provide
dried and crushed skins and seeds of the fruit (Tadeu-Pon- gaining valuable substances and at the same time reduce
tes, Carvalheiro, Roseiro, & Amaral-Collocßo, 1996). The the waste disposal problem.
skin, important component of pomace, is source of lyco- Dietary fiber has received increased attention recently.
pene. Lycopene is an excellent natural food color and also As consumers become more concerned about eating food
with health benefits, barley, which is naturally healthy, eas-
ily available and inexpensive crop is strongly favored for
Corresponding author. Tel.: +90 342 3172309; fax: +90 342 3601105. increased incorporation into human diet (Czuchajowska,
E-mail address: maskan@gantep.edu.tr (M. Maskan). Klamczynski, Paszczynska, & Baik, 1998). The dietary

0260-8774/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
232 A. Altan et al. / Journal of Food Engineering 84 (2008) 231–242

fiber content of barley contributes to its nutritional value, with mesh size of 20. Then, the sieved tomato pomace
making it a highly desirable cereal grain today. was finely ground and stored in polyethylene bags at
Extrusion cooking is an important and popular food 20 °C for further usage. The moisture content of dried
processing technique classified as a high temperature/short tomato pomace was 2.43 ± 0.2% (w.b.).
time process to produce fiber-rich products (Gaosong &
Vasanthan, 2000; Vasanthan, Gaosong, Yeung, & Li, 2.2. Sample preparation
2002). In the extruder, the food mix is thermomechanically
cooked to high temperature, pressure and shear stress Blends were prepared by mixing barley flour and tomato
which are generated in the screw-barrel assembly. The pomace in the ratios of 100:0, 98:2, 94:6, 90:10 and
cooked melt is then texturized and shaped in the die 87.27:12.73 on a dry-to-dry weight basis. These blends were
(Arhaliass, Bouvier, & Legrand, 2003). The thermome- chosen according to preliminary tests without jamming of
chanical action during extrusion brings about gelatiniza- extruder and for acceptable product’s physical characteris-
tion of starch, denaturation of protein and inactivation tics. The blended samples were conditioned to 21–22%
of enzymes, microbes and many anti-nutritional factors; (w.b.) moisture by spraying with a calculated amount of
all this occurs in a shear environment, resulting in a plasti- water and mixing continuously at medium speed in a mixer
cized continuous mass (Bhattacharya & Prakash, 1994). (Model F-30T, Blakeslee, Chicago, IL, USA). The samples
In recent years, there is an increasing demand for con- were put in buckets and stored at 4 °C overnight. The feed
version of fruit and vegetable wastes into useful products. material was then allowed 3 h to equilibrate at room tem-
The primary motivation is to minimize environmental perature prior to extrusion. This preconditioning proce-
impact of these by-products and to utilize valuable constit- dure was employed to ensure uniform mixing and
uents that remain, such as lycopene and dietary fiber. One hydration and to minimize variability in the state of the
viable method for utilization of fruit and vegetable by- feed material. Moisture content of samples was determined
products into useful products is extrusion processing due by halogen moisture analyzer (Model HR83 and HR83P,
to its versatility, high productivity, relative low cost, energy Mettler-Toledo GmbH, Greifensee, Switzerland) at 105 °C.
efficiency and lack of effluents. Successful incorporation of
tomato pomace into extruded products that deliver physi- 2.3. Extrusion cooking
ologically active components represents a major opportu-
nity for food processors providing the consumer a A laboratory-scale co-rotating twin-screw extruder
healthy barley-based product to choose from which is cur- (APV, Staffordshire, England) with a System9000 torque
rently lacking in the marketplace. Therefore, the objective rheometer (Haake Buchler, Paramus, NJ) that provided
of this research was to investigate processability of barley computer control and data acquisition was used. The slit
flour with the combination of tomato pomace to produce die (Haake Buchler, Paramus, NJ, USA) had dimensions
snack food in a twin-screw extruder. The effect of the vari- of 1.47 mm  20 mm  150 mm. The barrel diameter and
ables such as tomato pomace content, extrusion die tem- its length to diameter ratio (L/D) were 30 mm and 13:1,
perature and screw speed on system parameters and respectively. The MPC/V-30 had a clamshell barrel consist-
physical properties of extrudates were evaluated by using ing of three independent temperature zones controlled by
response surface methodology. Sensory properties were electrical heating and compressed air cooling. A computer-
determined in terms of color, texture, taste, off-odor and ized data acquisition system was used to control five set
overall acceptability for selected extrudate samples. temperatures and rotor speed and to record five melt tem-
peratures, pressure at the slit die and torque data. Data
2. Materials and methods acquisition rate was every 6 s. The barrel zone tempera-
tures were set at 30, 60, 100 and 130 °C throughout the
2.1. Materials experiments. The actual extruder screw speed is 2.5 times
the rotor speed. The screws were composed of screw ele-
Barley flour was obtained from Bob’s Red Mill Natural ments and lope-shaped paddles which could be assembled
Foods (Milwaukie, OR, USA). The particle size distribu- on the hexagon-shaped shafts to give different screw config-
tion of the barley flour was 12.1% (on mesh 40); 42.9% urations. The screw configuration used is shown in Fig. 1.
(on mesh 60); 38.9% (on mesh 80); 5.5% (on mesh 100); The screw configuration had three pieces of 1.5D twin lead
0.4% (on mesh 120) and 0.2% (mesh 120). Barley flour feed screws, two 1D twin lead feed screws, nine kneading
was stored at 4 °C until use. Tomato pomace, tomato-pro- elements oriented at 30° feed forward, one 1D single lead
cessing by-product, was obtained from the ConAgra Foods feed screw followed by nine kneading elements oriented
tomato processing plant located in Oakdale (California, at 30° feed forward and 1D discharge screw. Barley flour
USA). The pomace, obtained from the paste line, had a and tomato pomace blends were fed into extruder with a
moisture content of 46.4% (w.b.). It was dried at 50 °C K-tron Type T-20 twin-screw volumetric feeder (K-Tron
overnight in a forced-air drier (Model # R-4, Commercial Corp., Pitman, NJ, USA) at a rate of 2.11 ± 0.042 kg/h.
Dehydrator System, Inc., Eugene, OR, USA). The dried Extrudate was collected when the operation condition
tomato pomace was coarsely ground and passed on sieve was at steady state identified by torque value that vary less
A. Altan et al. / Journal of Food Engineering 84 (2008) 231–242 233

Fig. 1. Schematic representation of screw configuration.

than 5%. The samples were dried at 52 °C overnight in a formed for sensory data to determine differences between
forced-air drier (Model # R-4, Commercial Dehydrator treatments by using SPSS.
System, Inc., Eugene, OR, USA). The final dried samples
contained a maximum of 5.5% (w.b.) moisture. Dried sam- 2.5. System parameters
ples were stored in polyethylene bags at room temperature
and used for further analysis. Specific mechanical energy, the mechanical energy input
per unit mass of the extrudate, was calculated by dividing
2.4. Experimental design and data analysis the net power input to the screw by the extrudate flow rate.
SME input was calculated by the following equation
The central composite design for three independent vari- (Chang, Martinez-Bustos, Park, & Kokini, 1999; Fan,
ables was performed. The independent variables considered Mitchell, & Blanshard, 1996; Sokhey & Chinnaswamy,
were die temperature (X1), screw speed (X2) and pomace 1992):
level (X3). The independent variables and variation levels
screw speed ðs1 Þ  torque ðN mÞ
are shown in Table 1. The levels of each variable were SME ðWh kg1 Þ ¼ :
established according to literature data and preliminary tri- mass flow rate ðkg h1 Þ
als. The outline of experimental design with the coded and ð1Þ
actual levels is presented in Table 2. Dependent variables Torque was recorded every 6 s for at least 12 min and SME
were specific mechanical energy (SME), die melt tempera- was calculated and averaged for each processing condition.
ture, die pressure as system parameters and sectional Die pressure was measured using a Dynisco pressure trans-
expansion index (SEI), bulk density, water absorption ducer (PT-412, Dynisco, Franklin, MA, USA). Readings
and solubility indices, color and texture as product were recorded every 6 s for at least 12 min and average val-
responses. Response surface methodology was applied for ues were expressed as kPa. Die melt temperature was also
experimental data using a commercial statistical package,
Design-Expert version 6.0.6 (Statease Inc., Minneapolis,
MN, USA) for the generation of response surface plots. Table 2
The same software was used for statistical analysis of Experimental design for extrusion experiment with coded and actual
experimental data. The results were analyzed by a multiple variable levels
linear regression method which describes the effects of vari- Run Coded levels Actual levels
ables in first order, a two-factor interaction (2FI) and sec- X1 X2 X3 Die Screw Pomace
ond order polynomial models. Experimental data were temperature speed level (%)
fitted to the selected models and regression coefficients (°C) (rpm)
obtained. Statistical significance of the terms in the regres- 1 1 1 1 140 150 2
sion equation was examined by analysis of variance 2 1 1 1 160 150 2
(ANOVA) for each response. A Pearson’s correlation 3 1 1 1 140 200 2
matrix on product responses and system parameters was 4 1 1 1 160 200 2
5 1 1 1 140 150 10
carried out using SPSS 11.0 (SPSS Inc., Chicago, IL, 6 1 1 1 160 150 10
USA) in order to determine correlation coefficients 7 1 1 1 140 200 10
between parameters. Duncan’s multiple range test was per- 8 1 1 1 160 200 10
9 1.682 0 0 133.18 175 6
10 1.682 0 0 166.82 175 6
11 1 1.682 0 150 133 6
Table 1 12 1 1.682 0 150 217 6
Process variables used in the central composite design for three indepen- 13 0 0 1.682 150 175 0
dent variables 14 0 0 1.682 150 175 12.73
Code Variable level codes 15 0 0 0 150 175 6
16 0 0 0 150 175 6
1.682 1 0 1 1.682 17 0 0 0 150 175 6
Die temperature (°C) X1 133.18 140 150 160 166.82 18 0 0 0 160 175 6
Screw speed (rpm) X2 133 150 175 200 217 19 0 0 0 150 175 6
Pomace level (%) X3 0 2 6 10 12.73 20 0 0 0 150 175 6
234 A. Altan et al. / Journal of Food Engineering 84 (2008) 231–242

measured by thermocouple and monitored for every 6 s by 2.6.4. Texture

a computerized data acquisition system. The hardness of samples was measured with a TA-XT2i
Texture Analyzer (Texture Technologies Corp., Scarsdale,
2.6. Product responses NY, USA). Hardness in N was determined by measuring
the maximum force required to break the extruded samples
2.6.1. Expansion (42 mm long) using three point bend test with a sharp-
Expansion of extrudates was evaluated as sectional bladed probe (55 mm wide, 40 mm high, 9 mm thick).
expansion. The width and thickness of 15 pieces of extru- The test speed was 2 mm/s and the distance between two
date taken at random were measured with a digital caliper supports was 22 mm. A force–time curve was recorded
and the average calculated. The sectional expansion index and analyzed by Texture Exponent 32 software program
(SEI) was calculated using the equation proposed by Alva- (version 3.0). Elevan measurements were performed on
rez-Martinez, Kondury, and Harper (1988): each sample and averaged.
S e W e he
SEI ¼ ¼ ; ð2Þ 2.6.5. Color
S d W d hd
HunterLab LabScan XE (Hunter Associates Labora-
where Se and Sd are the cross-sectional areas of the extru- tory, Inc., Reston, VA, USA) was used to determine color
date and the die; We and he are the width and thickness of values of the raw materials and ground extruded in terms
the extrudate and Wd and hd are the width and thickness of of the L, a and b as measures of lightness, redness and yel-
the die, respectively. lowness, respectively. The measuring head was equipped
with 51 mm diameter viewing port and used the system
2.6.2. Bulk density of diffuse illumination with 10° viewing geometry. The
Bulk density was determined by measuring the volume illuminant was D65. The colorimeter was calibrated
of extrudate by glass bead displacement (Hwang & Hayak- against a standard white tile (L = 91.43, a = 0.74,
awa, 1980; Sokhey, Ali, & Hanna, 1997). Glass beads with b = 0.25). The extrudates were ground in a laboratory
a diameter range of 1.00–1.18 mm were used as displace- grinder and passed through a 60 mesh sieve prior to color
ment medium. Bulk densities of the extrudates were calcu- analysis. For each sample, four measurements were taken
lated as and averaged. The total color change (DE) was calculated
W ex as
qb ¼  qgb ; ð3Þ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
W gb 2 2 2
DE ¼ ðL  L0 Þ þ ðb  b0 Þ þ ða  a0 Þ ; ð4Þ
where qb is the bulk density using glass bead displacement
method (g/cm3), Wex is the extrudate mass (g), Wgb is the where the subscript ‘0’ indicates initial color values of the
mass of glass beads displaced (g) and qgb is the density of raw material.
the glass beads (g/cm3). The values were average of four
measurements. 2.7. Sensory evaluation

2.6.3. Water absorption and solubility indices A semi-trained panel of 34 students and faculty from
The water absorption index (WAI) is the weight of gel Food Engineering Department evaluated the extruded
obtained per gram of dry ground sample. The WAI of snacks for color, texture and overall acceptability on a 7-
extrudates was determined according to the AACC method point hedonic scale (from 1 = extremely dislike to
56–20 (AACC, 1995). The ground extrudate was suspended 7 = extremely like), while taste in terms of bran, tomato
in water at room temperature. After standing for 10 min, flavor and bitterness and off-odor was rated on a 7-point
gently stirred during this period, samples were centrifuged scale (from 1 = none to 7 = very high). Panelists rinsed
for 15 min at 1000g (AllegraTM 6 Centrifuge, Beckman their mouths with water after tasting each sample.
Coulter Inc., Palo Alto, CA, USA). The supernatant was
decanted into a tarred aluminum pan. The WAI was calcu- 3. Results and discussion
lated as the weight of sediment obtained after removal of
the supernatant per unit weight of original solids as dry Figures for die pressure, expansion, WAI, L, a and b
basis. The water solubility index (WSI) is the percentage were not given for the sake of simplicity.
of dry matter recovered after the supernatant is evaporated
from the water absorption determination. The supernatant 3.1. Diagnostic checking of fitted model and surface plots for
was dried in a vacuum oven at 84.4 °C and 20–24 mmHg various responses
gauge pressure for 24 h and weighed. The WSI was the
weight of dry solids in the supernatant expressed as a per- 3.1.1. Specific mechanical energy
centage of the original weight of sample on dry basis (Jin, A regression analysis were carried out to fit mathemati-
Hsieh, & Huff, 1995). WAI and WSI determinations were cal models to the experimental data. The predicted model
replicated four times.
A. Altan et al. / Journal of Food Engineering 84 (2008) 231–242 235

Table 3 effect of screw speed and tomato pomace level on SME is

Analysis of variance results for fitted models shown in Fig. 2. Increasing tomato pomace level in the
Response Source df Sum of Mean F- P-value blends increased the SME input in extrusion cooking sig-
squares squares value nificantly (P < 0.05). This effect could be explained by add-
SME Regression 9 46169.46 5129.94 13.37 0.0002* ing tomato pomace to barley flour gives a more viscous
Lack-of-fit 5 610.87 122.17 0.19 0.9541 melt requiring a higher torque and cause an increase in
Pure error 5 3225.64 645.13
Residual 10 3836.51 383.65
SME input. The observed effect of tomato pomace on
SME was similar to that reported by Hsieh, Huff, Lue,
Total 19 50005.97 and Stringer (1991) in extrusion of sugar beet fiber and
P Regression 6 3.820  107 6.367  106 15.07 <0.0001* corn meal. They reported that less water was available
Lack-of-fit 8 5.201  106 6.501  105 11.10 0.0084* for starches in corn meal in the presence of sugar beet fiber.
Pure error 5 2.928  105 58551.11 Because the viscosity of the starch-water system increases
Residual 13 5.493  106 4.226  105 with decreasing water content, torque and specific energy
increased with increasing sugar beet fiber. Statistical analy-
Total 19 4.370  107
sis revealed that SME was positively correlated with die
Significant at P < 0.05, df: degrees of freedom.
pressure (R = 0.564, P < 0.01) (Table 4). The measured
die melt temperature in extrusion cooking of barley flour
for specific mechanical energy (SME) can be described by and tomato pomace blends ranged from 129.55 to
the following equation in terms of coded values: 150.19 °C. Die melt temperature was negatively correlated
(R = 0.533, P < 0.05) with SME (Table 4). One might
SME ¼ 268:89  28:10X 1 þ 35:95X 2 þ 16:45X 3 expect that as the product temperature in the melting zone
increased, the viscosity of the dough would decrease which,
þ 14:01X 21 þ 11:50X 23  19:93X 1 X 3
in turn, would reduce torque and SME (Hsieh et al., 1991).
þ 17:03X 2 X 3 : ð5Þ Ryu and Ng (2001) reported that melt temperature in the
die exit affected SME input and decreased with the increase
The significance of coefficients of fitted quadratic model in melt temperature for both wheat flour and whole
(Eq. (5)) was evaluated by using the F-test and P-value. cornmeal.
Temperature (X1) had highly significant negative linear ef-
fect (P < 0.001) while screw speed (X2) and pomace level 3.1.2. Die pressure
(X3) had a significant positive linear effect on SME at The regression analysis results indicate that die pressure
P < 0.001 and P < 0.05 followed by a positive quadratic ef- (P) was highly significant (P < 0.001) on linear term of tem-
fect of temperature ðX 21 Þ (P < 0.05) and pomace level ðX 23 Þ perature (X1) and interaction term of temperature and
(P < 0.05). The interaction of temperature and pomace le- pomace level (X1X3). The regression equation obtained
vel (X1X3) had a significant negative effect (P < 0.05) for die pressure was as follows:
whereas the interaction of screw speed and pomace level
(X2X3) had a significant positive effect (P < 0.05) on SME. P ¼ 3806:71  1446:37X 1  1065:77X 1 X 3 : ð6Þ
The analysis of variance (ANOVA) for SME of qua-
dratic model (Eq. (5)) is given in Table 3. Regression model
fitted to experimental results of SME showed good correla-
tion coefficient (R2 = 0.9233). Table 3 shows that the F-
value for SME was significant (P < 0.05) whereas lack-of-
fit was not significant (P > 0.05). 334.68
SME (Wh/kg)

The measured SME in extrusion cooking of barley flour 305.46

and tomato pomace blends ranged from 163.37 to 276.24
372.13 W h/kg. SME decreased with increasing tempera- 247.02
ture and decreasing screw speed. Similar results were 217.80
observed by other authors (Doğan & Karwe, 2003; Koksel,
Ryu, Basman, Demiralp, & Ng, 2004; Ryu & Ng, 2001).
Increase in temperature suggests reduction in viscosity
which ultimately leads to reduced SME (Chang et al., 10.0
1999; Hsieh, Mulvaney, Huff, Lue, & Brent, 1989). An 200.0
increase in screw speed increased SME input. The increase 187.5
6.0 175.0
of SME with screw speed is evident from Eq. (5) which
Pomace level (%) 4.0 162.5 Screw speed (rpm)
shows that SME is proportional to the screw speed. Baik,
Powers, and Nguyen (2004) reported that increasing the 2.0 150.0
screw speed causes increases in SME input attributed to Fig. 2. Response surface plot for specific mechanical energy (SME) as a
the increase in shear rate with increased screw speed. The function of screw speed and pomace level at a temperature of 150 °C.
236 A. Altan et al. / Journal of Food Engineering 84 (2008) 231–242

Table 4
Correlation coefficients between product responses and system parameters
ns ns ns * * ns ns ns ns **
SEI 1 0.219 0.149 0.149 0.502 0.512 0.441 0.236 0.282 0.382 0.754 0.494*
BD 1 0.212ns 0.542* 0.106ns 0.113ns 0.180ns 0.376ns 0.925** 0.152ns 0.351ns 0.644**
WAI 1 0.184ns 0.508* 0.548* 0.543* 0.269ns 0.397ns 0.133ns 0.188ns 0.091ns
WSI 1 0.462* 0.433ns 0.476* 0.309ns 0.467* 0.782** 0.488* 0.618**
L 1 0.988** 0.974** 0.721** 0.031ns 0.271ns 0.089ns 0.296ns
a 1 0.992** 0.722** 0.038ns 0.212ns 0.080ns 0.296ns
b 1 0.752** 0.110ns 0.240ns 0.171ns 0.364ns
DE 1 0.319ns 0.131ns 0.235ns 0.323ns
H 1 0.134ns 0.444ns 0.637**
SME 1 0.564** 0.533*
P 1 0.777**
T 1
SEI: sectional expansion index; BD: bulk density; WAI: water absorption index; DE: total color change; H: hardness; SME: specific mechanical energy;
P: die pressure; T: die melt temperature.
Not significant.
Significant at P < 0.05.
Significant at P < 0.01.

The negative coefficient of the first order term of tempera- model fitted to experimental results of SEI showed higher
ture (X1) (Eq. (6)) indicated that die pressure increased with coefficient of determination (R2 = 0.9557). Table 5 shows
decrease of temperature. Meanwhile, negative coefficient of that the F-value for SEI was significant with a significant
interaction term (X1X3) of temperature and pomace level lack-of-fit (P < 0.05).
also resulted in decrease of die pressure. ANOVA for the The measured SEI of barley flour and tomato pomace
2FI model as fitted to experimental results (Table 3) shows blend extrudates varied between 0.893 and 2.014. When
significance (P < 0.05). The coefficient of determination extrusion-cooked melts exit the die, they suddenly go
(R2) for die pressure was 0.8743. Die pressure model from high pressure to atmospheric pressure. This pressure
showed significant (P < 0.05) lack-of-fit. The measured drop causes a flash-off of internal moisture and the water
die pressure in extrusion cooking of barley flour and toma- vapor pressure, which is nucleated to form bubbles in the
to pomace blends ranged from 786.45 to 6106.29 kPa. The molten extrudate, allows the expansion of the melt
pressure at the die exit was decreased upon increase in tem- (Arhaliass et al., 2003). SEI decreased when temperature
perature. Decrease in die pressure with the increase in tem- was increased. The lowest values for expansion were
perature may be attributed to decrease in viscosity of the found with temperature of 166.8 °C. Expansion decrease
melt (Ryu & Ng, 2001; Singh & Smith, 1997) due to degra- at higher extruder temperatures can be attributed to
dation of gelatinized starch granules (Cai, Diosady, & Ru- increase dextrinization and weakening of structure (Men-
bin, 1995; Singh, Sekhon, & Singh, 2007). A negative doncßa, Grossmann, & Verhé, 2000). Launay and Lisch
correlation was found between die pressure and melt tem- (1983) proposed that the corn extrudate longitudinal
perature (R = 0.777, P < 0.01) (Table 4). It was observed and diametral (sectional) expansions depended on the
that increasing pomace level with increasing temperature melt viscosity and elasticity. They reported that an
decreased die pressure. increased water content or temperature would yield a
lower melt viscosity and increased longitudinal expansion
3.1.3. Expansion while the melt elasticity would be lowered and a decrease
The regression equation for expansion as sectional in diametral expansion would be observed. This result is
expansion index (SEI) at any temperature (X1) and pomace in agreement also with the works of other researchers
level (X3) was (Doğan & Karwe, 2003; Ilo, Liu, & Berghofer, 1999).
Several researchers have demonstrated that the expansion
SEI ¼ 1:59  0:25X 1  0:18X 3  0:14X 21 þ 0:073X 23 ratio of extruded cereals depends on the degree of starch
þ 0:099X 1 X 3 : ð7Þ gelatinization (Case, Hanna, & Scwartz, 1992; Chinnasw-
amy & Hanna, 1988). However, increasing level of
It was observed that temperature (X1) and pomace level tomato pomace resulted in decrease in SEI of extrudates.
(X3) had highly significant negative linear effect This may be attributed to dilution effect of pomace on
(P < 0.001) on SEI followed by a negative quadratic effect starch. Screw speed had no significant effect (P > 0.05)
of temperature ðX 21 Þ (P < 0.001) and a positive quadratic on expansion of extrudates. Increasing level of pomace
effect of pomace level ðX 23 Þ (P < 0.05). The interaction of with increasing temperature decreased expansion. Sec-
temperature and pomace level (X1X3) had a significant po- tional expansion index was correlated with die melt tem-
sitive effect (P < 0.05) on SEI (Eq. (7)). ANOVA for qua- perature (R = 0.494, P < 0.05) and pressure (R = 0.754,
dratic model of SEI is given in Table 5. Regression P < 0.001). Sokhey et al. (1997) concluded that radial
A. Altan et al. / Journal of Food Engineering 84 (2008) 231–242 237

Table 5
Analysis of variance results for fitted models
Response Source df Sum of squares Mean squares F-value P-value
SEI Regression 9 1.71 0.19 23.95 <0.0001*
Lack-of-fit 5 0.071 0.014 9.00 0.0154*
Pure error 5 7.913  103 1.583  103
Residual 10 0.079 7.914  103

Total 19 1.78
BD Regression 9 0.50 0.055 13.58 0.0002*
Lack-of-fit 5 0.033 6.611  103 4.29 0.0679
Pure error 5 7.699  103 1.540  103
Residual 10 0.041 4.075  103

Total 19 0.54
WAI Regression 3 0.76 0.25 6.70 0.0039*
Lack-of-fit 11 0.46 0.041 1.40 0.3741
Pure error 5 0.15 0.030
Residual 16 0.60 0.038

Total 19 1.36
WSI Regression 3 44.94 14.98 26.23 <0.0001*
Lack-of-fit 11 8.47 0.77 5.74 0.0333*
Pure error 5 0.67 0.13
Residual 16 9.14 0.57

Total 19 54.08
H Regression 9 742.88 82.54 87.91 <0.0001*
Lack-of-fit 5 9.34 1.87 176.61 <0.0001*
Pure error 5 0.053 0.011
Residual 10 9.39 0.94

Total 19 752.27
Significant at P < 0.05, df: degrees of freedom.

expansion (sectional expansion) depended directly on the expansion only in the direction perpendicular to extrudate
pressure at the die nozzle that correlates our results. flow, while unit bulk density considers expansion in all
directions (Falcone & Phillips, 1988). The bulk density ran-
3.1.4. Bulk density ged from 0.370 to 1.111 g/cm3 for the barley flour–tomato
The quadratic model obtained from regression analysis pomace extrudates. The lowest bulk density value was
for bulk density (BD) in terms of coded levels of the vari- obtained at higher temperatures with a low level of tomato
ables was developed as follows: pomace, whereas the highest value was obtained at lower
temperatures (Fig. 3). A negative influence of temperature
BD ¼ 0:53  0:15X 1 þ 0:093X 21 þ 0:085X 1 X 3 : ð8Þ
was found on bulk density of barley flour and tomato pom-
Bulk density of barley flour and tomato pomace extrudate ace extrudate, which agrees with other investigations in lit-
was significantly affected (P < 0.001) by the linear and qua- erature (Doğan & Karwe, 2003; Hagenimana, Ding, &
dratic terms of temperature (X1 and X 21 ) but was not signif- Fang, 2006; Ilo et al., 1999). Screw speed had no effect
icantly (P > 0.05) dependent on screw speed (X2) and on bulk density of extrudates. The response surface graph
pomace level (X3). The interaction term of temperature shows that bulk density decreased with increasing temper-
and pomace level (X1X3) was found to be significant ature, whereas increased with an increase in level of pom-
(P < 0.001). ANOVA for bulk density of quadratic model ace as shown in Fig. 3. Bulk density was negatively
(Eq. (8)) is given in Table 5. Regression model fitted to correlated (R = 0.644, P < 0.01) with die melt tempera-
experimental results of bulk density showed good correla- ture (Table 4).
tion coefficient (R2 = 0.9244). Table 5 shows that the F-va-
lue for bulk density was significant (P < 0.05), whereas 3.1.5. Water absorption and solubility indices
lack-of-fit was not significant (P > 0.05). Multiple regression equations were generated relating to
The expansion ratio and bulk density of extrudates seek water absorption index (WAI) and water solubility index
to describe the degree of puffing undergone by the dough as (WSI) to coded levels of the variables. Model was devel-
it exits the extruder. Sectional expansion index considers oped as follows:
238 A. Altan et al. / Journal of Food Engineering 84 (2008) 231–242

holding capacity when the ratio of fiber/corn starch

increased in extrusion of corn fiber and corn starch blend.
Bulk density (g/cm3)

In addition, Singh et al. (2007) observed a decrease in WAI
with addition of pea grits in extrusion of rice. They
explained that a decrease in WAI was due to the dilution
of starch in rice pea blends.
0.484 The WSI ranged from 7.08% to 12.99% for the barley
0.367 flour–tomato pomace extrudates. The effect of temperature
and screw speed on WSI of extrudates is shown in Fig. 4.
The WSI increased significantly (P < 0.001) with increasing
screw speed and tomato pomace level and decreasing tem-
140.0 10.0 perature. The increase in WSI with increasing screw speed
145.0 8.0 was consistent with the results reported for corn meal and
150.0 6.0 corn and wheat extrudates (Jin et al., 1995; Mezreb, Goul-
4.0 Pomace level (%) lieux, Ralainirina, & Queneudec, 2003). Mezreb et al.
Temperature (º C) 155.0
160.0 2.0 (2003) reported that the increase of screw speed induced
a sharp increase of specific mechanical energy, the high
Fig. 3. Response surface plot for bulk density (BD) as a function of
pomace level and temperature at a screw speed of 175 rpm.
mechanical shear degraded macromolecules, and so the
molecular weight of starch granules decreased and hence
increased WSI. In this study, it was observed that WSI
WAI ¼ 6:54  0:12X 1  0:18X 3 ; ð9Þ was positively correlated with SME (R = 0.782, P < 0.01),
die pressure (R = 0.488, P < 0.05) and bulk density
WSI ¼ 9:66  1:15X 1 þ 1:03X 2 þ 0:98X 3 : ð10Þ
(R = 0.542, P < 0.05) but negatively correlated with die
The negative coefficients of the linear terms of temperature melt temperature (Table 4). Temperature was found being
(X1) and pomace level (X3) (Eq. (9)) indicated that WAI de- inversely proportional to WSI; that is, the higher the extru-
creases with increase of these variables while positive coef- sion temperatures, the lower the WSI values. Similar find-
ficients (Eq. (10)) of the linear terms of screw speed (X2) ings were achieved by Gutkoski and El-Dash (1999) in
and pomace level (X3) resulted increase in WSI. ANOVA extruded oat products.
for models of WAI and WSI is given in Table 5. As indi-
cated in ANOVA table, first order model for WAI and 3.1.6. Texture
WSI was found to be significant (P < 0.05). However, the The quadratic model for hardness (H) in terms of coded
lack-of-fit was not significant (P > 0.05) for WAI but sig- levels of the variables was developed as follows:
nificant for WSI (P < 0.05). The coefficients of determina-
H ¼ 7:70  5:79X 1  0:67X 2 þ 4:13X 21 þ 1:93X 1 X 3 : ð11Þ
tion (R2) for water absorption and solubility indices were
0.5569 and 0.8310, respectively. Hardness of the barley flour and tomato pomace extrudate
The WAI measures the volume occupied by the granule was significantly affected by linear terms of temperature
or starch polymer after swelling in excess water. While the
WSI determines the amount of free polysaccharide or poly-
saccharide released from the granule after addition of
excess water (Sriburi & Hill, 2000). The WAI ranged from
6.10 to 7.03 g/g for the barley flour–tomato pomace extru-
dates. Increasing temperature significantly (P < 0.05)
WSI (%)

decreased the WAI of extrudates. Similar results were

reported by Guha, Ali, and Bhattacharya (1997), Pelembe, 9.66
Erasmus, and Taylor (2002) Ding, Ainsworth, Plunkett, 8.57
Tucker, and Marson (2006). A decrease in WAI with 7.48
increasing temperature was probably due to decomposition
or degradation of starch (Pelembe et al., 2002). Ding et al.
(2006) also stated that the WAI decreases with increasing
temperature if dextrinization or starch melting prevails 200.0
over the gelatinization phenomenon. The WAI decreased 187.5 160.0
significantly (P < 0.01) as the percentage of tomato pomace 155.0
increased. This may be attributed to relative decrease in 150.0
Screw speed (rpm) 162.5 145.0
starch content with addition of pomace and competition Temperature (ºC)
of absorption of water between pomace and available 150.0 140.0
starch. This result is in agreement with those of Artz, War- Fig. 4. Response surface plot for water solubility index (WSI) as a
ren, and Villota (1990). They reported a decrease in water function of temperature and screw speed at a pomace level of 6%.
A. Altan et al. / Journal of Food Engineering 84 (2008) 231–242 239

(X1) and screw speed (X2) at P < 0.001 and P < 0.05, Table 6
respectively. Temperature had also significant quadratic ef- Analysis of variance results for fitted models
fect (P < 0.001) on hardness of extrudates. The interaction Response Source df Sum of Mean F- P-value
term (X1X3) between temperature and pomace level was squares squares value
significant, so that high values of hardness were found at L-value Regression 9 139.17 15.46 36.91 <0.0001*
high level of pomace, dependent on temperature. Regres- Lack-of-fit 5 1.98 0.40 0.90 0.5468
Pure error 5 2.21 0.44
sion model (Eq. (11)) fitted to experimental results of hard- Residual 10 4.19 0.42
ness showed higher coefficient of determination
(R2 = 0.9875). Table 5 shows that the F-value for hardness Total 19 143.36
was significant with a significant lack-of-fit (P < 0.05). a-Value Regression 9 138.73 15.41 218.42 <0.0001*
The textural property of barley flour and tomato pom- Lack-of-fit 5 0.42 0.085 1.50 0.3335
ace extrudate was determined by measuring the force Pure error 5 0.28 0.056
required to break the extrudate (Singh, Hoseney, & Fau- Residual 10 0.71 0.071
bion, 1994). The higher the value of maximum peak force
Total 19 139.44
required in gram, which means the more force required
to breakdown the sample, the higher the hardness of the b-Value Regression 9 113.23 12.58 154.03 <0.0001*
Lack-of-fit 5 0.68 0.14 5.07 0.0497*
sample to fracture (Li, Zhang, Tony Jin, & Hsieh, 2005).
Pure error 5 0.13 0.027
The effect of temperature and tomato pomace level on Residual 10 0.82 0.082
hardness of extrudates is shown in Fig. 5. Response surface
plot showed that a decrease in die temperature with Total 19 114.05
increasing level of tomato pomace increased the product DE Regression 9 15.57 1.73 6.85 0.0029*
hardness. Hardness of barley flour and tomato pomace Lack-of-fit 5 1.20 0.24 0.91 0.5409
extrudate varied between 5.64 and 29.75 N. A decrease in Pure error 5 1.32 0.26
die temperature increased the product hardness giving a Residual 10 2.52 0.25
maximum at about 133.18 °C, 175 rpm screw speed and Total 19 18.09
6% tomato pomace level. This result is in line with density *
Significant at P < 0.05, df: degrees of freedom.
where an increase in density was observed. High density
product naturally offers high hardness evident by high cor-
relation between product density and hardness (R = 0.925,
P < 0.01) (Table 4). Similar effect of temperature has been 3.1.7. Color
observed while extruding yam and wheat flour (Ding et al., Color is an important quality factor directly related to
2006; Sebio & Chang, 2000). Increasing screw speed the acceptability of food products, and is an important
slightly decreased the hardness of the barley flour–tomato physical property to report for extrudate products. The
pomace extrudate, particularly at higher temperatures. regression equations for color parameters (L, a and b val-
Liu, Hsieh, Heymann, and Huff (2000) found that the ues) and total color change (DE) at any temperature (X1)
hardness of the extruded oat–corn flour increased as the and pomace level (X3) were
screw speed decreased.
L ¼ 69:95  3:23X 3 þ 0:79X 23 ; ð12Þ
a ¼ 9:88 þ 3:21X 3  0:92X 23 ; ð13Þ
b ¼ 22:89  0:32X 1 þ 2:88X 3 þ 0:17X 21  0:86X 23 ; ð14Þ
DE ¼ 8:90  0:31X 1 þ 0:69X 3  0:80X 23 : ð15Þ
Hardness (N)

16.33 Tomato pomace level (X3) was an important variable in the

12.47 response surface models (Eqs. (12)–(15)) of product color
parameters, as its linear and quadratic terms were signifi-
cant at P < 0.01 and P < 0.001, respectively. The color
4.74 parameter b of barley flour and tomato pomace extrudate
was significantly (P < 0.01, P < 0.05) affected by linear
and quadratic terms of temperature (X1). Temperature
had also significant (P < 0.05) effect on DE of extrudates.
140.0 10.0 ANOVA results for models of L, a and b color parameters
145.0 8.0 and DE are given in Table 6. As indicated in ANOVA ta-
150.0 6.0 ble, quadratic model for L, a, b color parameters and total
Temperature (ºC) 155.0 4.0 Pomace level (%)
color change, DE was found to be significant (P < 0.05).
160.0 2.0 However, the lack-of-fit was not significant (P > 0.05) for
Fig. 5. Response surface plot for hardness (H) as a function of pomace L, a color parameters and DE but significant for b color
level and temperature at a screw speed of 175 rpm. parameter (P < 0.05). The coefficients of determination
240 A. Altan et al. / Journal of Food Engineering 84 (2008) 231–242

(R2) for L, a, b color parameters and DE were 0.9708, Table 7

0.9949, 0.9928 and 0.8605, respectively. Sensory evaluation scores of extrudates produced at different conditions
The non-extruded blend of barley flour and tomato Extrudates
pomace with a percentage of 0, 2, 6, 10 and 12.73 pomace AA B C D E
had color values of the ranges: L: 70.79–79.17; a: 2.02–9.22; Color 3.64a 4.70bc 4.85bc 5.08c 4.23ab
b: 11.38–19.09, whereas the barley flour–tomato pomace Texture 3.91a 5.32b 4.58c 5.26b 3.44a
extrudates had color values of the ranges: L, 65.85–76.32; Taste
a, 2.70–12.66; b, 16.10–25.63. Among the color parameters, Bran flavor 3.67a 3.79a 3.23a 3.44a 3.82a
the L and a values showed marked changes due to addition Tomato flavor 1.67a 1.50a 2.70b 2.50b 3.02b
of tomato pomace only. An increase in tomato pomace Bitterness 1.85a 1.61a 1.94a 1.88a 2.00a
level decreased the L value of the samples and increased Off-odor 1.79a 1.79a 1.70a 1.58a 1.88a
the a value of samples as expected due to the lycopene pig- Overall acceptability 3.94a 4.85b 4.94b 5.23b 4.08a
ment in the tomato pomace. Statistical analysis showed A: 0% pomace level, 150 °C, 175 rpm; B: 2% pomace level, 160 °C,
that there was a negative correlation between L and a val- 200 rpm; C: 6% pomace level, 150 °C, 217 rpm; D: 10% pomace level,
ues of samples (R = 0.988, P < 0.01) (Table 4). Similar 160 °C, 200 rpm; E: 12.73%, 150 °C, 175 rpm.
Means within a row with different superscripts are significantly dif-
result was found by Ilo and Berghofer (1999). The L value
ferent (P < 0.05).
was positively correlated with SEI (R = 0.502, P < 0.05)
and WAI (R = 0.508, P < 0.05) but negatively correlated
with WSI (R = 0.462, P < 0.05). On the other hand, a Total color change in extruded products ranged between
value was negatively correlated with SEI (R = 0.512, 5.56 and 9.99. The effect of temperature and pomace level
P < 0.05) and WAI (R = 0.548, P < 0.05). An increase on total color change of extrudates is shown in Fig. 6.
in expansion gives more bright color in extrudates due to Results of regression analysis show that color change was
air cells rather than dull color. The change in yellowness most dependent on tomato pomace content (P < 0.001).
(b value) decreased with increasing temperature which is It was observed that DE was negatively correlated with L
in agreement with the results of Ilo and Berghofer (1999). value (R = 0.721, P < 0.01) and positively correlated with
They reported that the changes in yellowness during extru- a value (R = 0.722, P < 0.01) and b value (R = 0.752,
sion cooking of yellow maize induced by the effects of two P < 0.01) (Table 4).
different reactions: the non-enzymatic browning and pig-
ment destruction. They also concluded that some of the 3.2. Sensory evaluation
caratenoids might have been damaged by the thermal treat-
ment and some browning might have made up the color Five extrudate samples were selected out of 20 extrudate
loss. Increasing tomato pomace content resulted in a signif- samples with respect to textural property and different pom-
icant (P < 0.001) increase in the extrudate b value and DE. ace level for sensory evaluation. The mean values of sensory
The b value was positively correlated with a value panel ratings of extrudates are presented in Table 7. Extru-
(R = 0.992, P < 0.01) and WSI (R = 0.476, P < 0.05), dates with different level of tomato pomace had better score
whereas negatively correlated with L value (R = 0.974, than that of extrudate with 0%. Extrudate D with 10%
P < 0.01) and WAI (R = 0.543, P < 0.05) (Table 4). tomato pomace had the highest level of acceptance for
color. Extrudates B and D had higher preference values
for the parameter of texture. There were no significant dif-
ferences (P > 0.05) in bran flavor, bitterness and off-odor
scores among extrudates. Tomato flavor score changed as
9.61 a result of increasing percentage of tomato pomace. How-
9.04 ever, tomato flavor was perceived as weak (3.02) by panel-
ists for highest level of pomace. The overall acceptability of

7.91 the barley flour and tomato pomace extrudate ranged low-
est (3.94) in extrudate A and highest (5.23) in extrudate D.

4. Conclusion

The system parameters and product responses were

140.0 10.0 found to be most dependent on temperature and pomace
145.0 8.0 level. The results showed that varying levels of tomato
150.0 6.0 pomace could be incorporated into an extruded barley
Temperature ( C)
º 155.0 4.0 Pomace level (%)
snack depending on the desired texture of the final product.
160.0 2.0 Extrudates with 2% and 10% tomato pomace levels
Fig. 6. Response surface plot for total color change (DE) as a function of extruded at 160 °C and 200 rpm had higher preference lev-
pomace level and temperature at a screw speed of 175 rpm. els for parameters of color, texture, taste and overall
A. Altan et al. / Journal of Food Engineering 84 (2008) 231–242 241

acceptability. Such extrusion would also provide another properties of snack-like products prepared from cowpea and sorghum
avenue for tomato pomace utilization. flours by extrusion. Cereal Chemistry, 53, 1464–1469.
Fan, J., Mitchell, J. R., & Blanshard, J. M. V. (1996). The effect of sugars
on the extrusion of maize grits: I. The role of the glass transition in
Acknowledgements determining product density and shape. International Journal of Food
Science and Technology, 31, 55–65.
Gaosong, J., & Vasanthan, T. (2000). The effect of extrusion cooking on
This research project was supported by the University of the primary structure and water solubility of b-glucans from regular
Gaziantep (Turkey) and Center for Advanced Materials, and waxy barley. Cereal Chemistry, 77, 396–400.
Methods and Processing, supporting the mission of the Guha, M., Ali, S. Z., & Bhattacharya, S. (1997). Twin-screw extrusion of
Robert Mondavi Institute for Wine and Food Science, Uni- rice flour without die: Effect of barrel temperature and screw speed on
versity of California, Davis. Special thanks are extended to extrusion and extrudate characteristics. Journal of Food Engineering,
32, 251–267.
Mr. Jannes Vandeven for assistance with extrusion process- Gutkoski, L. C., & El-Dash, A. A. (1999). Effect of extrusion process
ing and Mr. Ken Shaw, principal mechanician, Department variables on physical and chemical properties of extruded oat
of Food Science and Technology, UC Davis. products. Plant Foods for Human Nutrition, 54, 315–325.
Hagenimana, A., Ding, X., & Fang, T. (2006). Evaluation of rice
flour modified by extrusion cooking. Journal of Cereal Science,
References 43, 38–46.
Hsieh, F., Huff, H. E., Lue, S., & Stringer, L. (1991). Twin screw extrusion
AACC (1995). Approved methods of American association of cereal of sugar-beet fiber and corn meal. Lebensmitted-Wissenschaft Unter-
chemists (9th ed.). Methods 56–20. The Association: St. Paul, MN. suchung-Technology, 24, 495–500.
Alvarez-Martinez, L., Kondury, K. P., & Harper, J. M. (1988). A general Hsieh, F., Mulvaney, S. J., Huff, H. E., Lue, S., & Brent, J. (1989). Effect
model for expansion of extruded products. Journal of Food Science, 53, of dietary fiber and screw speed on some extrusion processing and
609–615. product variables. Lebensmitted-Wissenschaft Untersuchung-Technol-
Arhaliass, A., Bouvier, J. M., & Legrand, J. (2003). Melt growth and ogy, 22, 204–207.
shrinkage at the exit of the die in the extrusion-cooking process. Hwang, M. P., & Hayakawa, K. I. (1980). Bulk densities of cookies
Journal of Food Engineering, 60, 185–192. undergoing commercial baking processes. Journal of Food Science, 45,
Artz, W. E., Warren, C. C., & Villota, R. (1990). Twin screw extrusion 1400–1402.
modification of corn fiber and corn starch extruded blend. Journal of Ilo, S., & Berghofer, E. (1999). Kinetics of colour changes during
Food Science, 55, 746–750, 754. extrusion cooking of maize gritz. Journal of Food Engineering, 39,
Baik, B. K., Powers, J., & Nguyen, L. T. (2004). Extrusion of regular and 73–80.
waxy barley flours for production of expanded cereals. Cereal Ilo, S., Liu, Y., & Berghofer, E. (1999). Extrusion cooking of rice flour and
Chemistry, 81, 94–99. amaranth blends. Lebensmitted-Wissenschaft Untersuchung-Technol-
Bhattacharya, S., & Prakash, M. (1994). Extrusion of blends of rice and ogy, 322, 79–88.
chick pea flours: A response surface analysis. Journal of Food Jin, Z., Hsieh, F., & Huff, H. E. (1995). Effects of soy fiber, salt, sugar, and
Engineering, 21, 315–330. screw speed on physical properties and microstructure of corn meal
Bobek, P., Ozdin, L., & Hromadova, M. (1998). The effect of dried tomato extrudate. Journal of Cereal Science, 22, 185–194.
grape and apple pomace on the cholesterol metabolism and antiox- Johnson, E. J. (2000). The role of lycopene in health and disease. Nutrition
idative enzymatic system in rats with hypercholesterolemia. Nahrung, in Clinical Care, 3, 35–43.
42, 317–320. Kaur, D., Sogi, D. S., Gary, S. K., & Bawa, A. S. (2005). Flotation-cum-
Cai, W., Diosady, L. L., & Rubin, L. J. (1995). Degradation of wheat sedimentation system for skin and seed separation from tomato
starch in a twin-screw extruder. Journal of Food Engineering, 26, pomace. Journal of Food Engineering, 71, 341–344.
289–300. Koksel, H., Ryu, G. H., Basman, A., Demiralp, H., & Ng, P. K. W.
Case, S. E., Hanna, M. A., & Scwartz, S. J. (1992). Effect of starch (2004). Effects of extrusion variables on the properties of waxy hulless
gelatinization on physical properties of extruded wheat and corn based barley extrudates. Nahrung, 48, 19–24.
products. Cereal Chemistry, 69, 401–404. Launay, B., & Lisch, L. M. (1983). Twin-screw extrusion cooking of
Chang, Y. K., Martinez-Bustos, F., Park, T. S., & Kokini, J. L. (1999). starches: Flow behaviour of starch pastes, expansion and
The influence of specific mechanical energy on cornmeal viscosity mechanical properties of extrudates. Journal of Food Engineering,
measured by an on-line system during twin-screw extrusion. Brazilian 2, 259–280.
Journal of Chemical Engineering, 16, 285–295. Li, S., Zhang, H. Q., Tony Jin, Z., & Hsieh, F. (2005). Textural
Chinnaswamy, R., & Hanna, M. A. (1988). Relationship between amylose modification of soya bean/corn extrudates as affected by moisture
content and extrusion-expansion properties of corn starches. Cereal content, screw speed and soya bean concentration. International
Chemistry, 65, 138–143. Journal of Food Science and Technology, 40, 731–741.
Czuchajowska, Z., Klamczynski, A., Paszczynska, B., & Baik, B.-K. Liu, Y., Hsieh, E., Heymann, H., & Huff, H. E. (2000). Effect of process
(1998). Structure and functionality of barley starches. Cereal Chem- conditions on the physical and sensory properties of extruded oat–corn
istry, 75, 747–754. puff. Journal of Food Science, 65, 1253–1259.
Del Valle, M., Camara, M., & Torija, M.-E. (2006). Chemical character- Mendoncßa, S., Grossmann, M. V. E., & Verhé, R. (2000). Corn bran as a
ization of tomato pomace. Journal of the Science of Food and fibre source in expanded snacks. Lebensmitted-Wissenschaft Untersu-
Agriculture, 86, 1232–1236. chung-Technology, 33, 2–8.
Ding, Q. B., Ainsworth, P., Plunkett, A., Tucker, G., & Marson, H. Mezreb, K., Goullieux, A., Ralainirina, R., & Queneudec, M. (2003).
(2006). The effect of extrusion conditions on he functional and physical Application of image analysis to measure screw speed influence on
properties of wheat-based expanded snacks. Journal of Food Engineer- physical properties of corn and wheat extrudates. Journal of Food
ing, 73, 142–148. Engineering, 57, 145–152.
Doğan, H., & Karwe, M. V. (2003). Physicochemical properties of quinoa Pelembe, L. A. M., Erasmus, C., & Taylor, J. R. N. (2002). Development
extrudates. Food Science and Technology International, 9, 101–114. of a protein-rich composite sorghum-cowpea instant porridge by
Falcone, R. G., & Phillips, R. D. (1988). Effects of feed composition, feed extrusion cooking process. Lebensmitted-Wissenschaft Untersuchung-
moisture and barrel temperature on the physical and rheological Technology, 35, 120–127.
242 A. Altan et al. / Journal of Food Engineering 84 (2008) 231–242

Ryu, G. H., & Ng, P. K. W. (2001). Effects of selected process parameters Sokhey, A. S., & Chinnaswamy, R. (1992). Physicochemical properties
on expansion and mechanical properties of wheat flour and whole of irradiation modified starch extrudates. Food Structure, 11,
cornmeal extrudates. Starch, 53, 147–154. 361–371.
Sebio, L., & Chang, Y. K. (2000). Effects of selected process parameters in Sokhey, A. S., Ali, Y., & Hanna, M. A. (1997). Effects of die dimensions
extrusion of yam flour (Dioscorea rotundata) on physicochemical on extruder performance. Journal of Food Engineering, 31, 251–261.
properties of extrudates. Nahrung, 44, 96–100. Sriburi, P., & Hill, S. E. (2000). Extrusion of cassava starch with either
Singh, B., Sekhon, K. S., & Singh, N. (2007). Effects of moisture, variations in ascorbic acid concentration or pH. International Journal
temperature and level of pea grits on extrusion behaviour and product of Food Science and Technology, 35, 141–154.
characteristics of rice. Food Chemistry, 100, 198–202. Tadeu-Pontes, M. A., Carvalheiro, F., Roseiro, J. C., & Amaral-Collocßo,
Singh, J., Hoseney, R. C., & Faubion, J. M. (1994). Effect of dough M. T. (1996). Evaluation of product composition profile during an
properties on extrusion-formed and baked snacks. Cereal Chemistry, extrusion based process of tomato pomace transformation. Agro-Food-
71, 417–422. Industry Hi-Tech, May/June, 39–40.
Singh, N., & Smith, A. C. (1997). A comparison of wheat starch, whole Vasanthan, T., Gaosong, J., Yeung, J., & Li, J. (2002). Dietary fiber
wheat meal and oat flour in the extrusion cooking process. Journal of profile of barley flour as affected by extrusion cooking. Food
Food Engineering, 34, 15–22. Chemistry, 77, 35–40.