Journal of Food Engineering 90 (2009) 262–270

Contents lists available at ScienceDirect

Journal of Food Engineering journal homepage: www.elsevier.com/locate/jfoodeng

Soy protein-fortified expanded extrudates: Baseline study using normal corn starch Normell Jhoe E. de Mesa a,*, Sajid Alavi a, Narpinder Singh b, Yong-Cheng Shi a, Hulya Dogan a, Yijun Sang a a b

Department of Grain Science and Industry, Kansas State University, 201 Shellenberger Hall, Manhattan, Kansas 66506, United States Department of Food Science and Technology, Guru Nanak Dev University, Amritsar-143 005, India

a r t i c l e

i n f o

Article history: Received 21 December 2007 Received in revised form 13 June 2008 Accepted 21 June 2008 Available online 3 July 2008 Keywords: Starch–protein interactions Cellular microstructure Extrudate texture

a b s t r a c t Soy protein supplementation increases the nutritional value of starch-based expanded snacks. A systematic study was conducted to serve as baseline for optimizing the addition of soy protein concentrate (SPC). Physical and microstructural properties of native corn starch–soy protein concentrate (CS–SPC) extrudates were investigated in relation to the macromolecular changes in starch during extrusion. The effects of extruder screw speed (230 and 330 rpm) and SPC concentration (0%, 5%, 10%, 15%, 20%) on the abovementioned parameters were determined. Increasing screw speed resulted in higher specific mechanical energy (SME) and expansion, and lower mechanical strength. On the other hand, addition of 5–20% SPC led to lower SME and expansion, and higher mechanical strength. X-ray micrographs showed smaller yet more cells, and cell wall thickening with SPC addition. Water absorption index increased and water solubility index decreased with increase in screw speed and SPC level. Increasing screw speed resulted in a slight shift towards smaller molecular weight fractions of starch, as determined by gel permeation chromatography. Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction Soy protein is widely used in food applications due to its functionality and health benefits (Liu, 1997; Riaz, 2006). Effective October 1999, the US Food and Drug Administration has approved the use of soy protein health claims on food labels based on human intervention studies and clinical trials that show a high association between consumption of soy protein and the reduced risk of coronary heart disease (e-CFR 101.82, 1999). In addition to the cholesterol lowering effect of soy protein, it has anticarcinogenic effects, and it deters obesity, diabetes, digestive tract irritation, and bone and kidney diseases (Friedman and Brandon, 2001; Messina and Barnes, 1991). While starch is the primary ingredient in expanded breakfast cereals and snacks, soy protein can enhance the nutritional value of these products. Incorporation of soy protein can significantly impact the mechanical, physico-chemical and microstructural properties of foods. Several studies have investigated the addition of soy protein to extruded starch products with conflicting results (Chang et al., 2001; Faubion and Hosney, 1982; Ghorpade et al., 1997; Li et al., 2007; Zasypkin and Lee, 1998). Chang et al. (2001) found that radial expansion increased on addition of up to 25% SPC to a cassava starch-based matrix. Further addition of SPC decreased the radial expansion. Additionally, with increasing soy pro-

* Corresponding author. Tel.: +17855324073; fax: +17855324017. E-mail address: [email protected] (N.J.E. de Mesa). 0260-8774/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jfoodeng.2008.06.032

tein concentrate, hardness decreased, water absorption index (WAI) increased, and water solubility index (WSI) decreased. Faubion and Hosney (1982) reported that the expansion of wheat starch with 1–8% soy protein isolate was higher than that of pure starch. However, at 10% soy protein isolate, expansion decreased. Ghorpade et al. (1997), on the other hand, showed that increasing the percentage of soy protein isolate from 10% to 30% in corn starch extrudates did not significantly affect bulk density and the percentage of open pores. Zasypkin and Lee (1998) showed that increasing the proportion of soybean flour to 10% in a wheat flour-soybean flour blend resulted in a decrease in expansion ratio (ER) at 16% in-barrel moisture content (MC), and an increase in ER at 17–18% MC. Further addition of soy flour up to 40% led to continuous decrease in expansion ratio. Beyond 40%, the trends for ER were again dependent on MC and also a possible phase inversion occurring beyond that level of soy flour. Li et al. (2005) reported that the addition of soybean flour to corn meal in the range of 0 to 40% increased ER. They also found significant interactions between MC and soy flour levels. In addition to ingredient composition, processing parameters, such as MC and screw speed, also affect extrudate properties. Effect of MC has already been discussed above. By modeling the behavior of ER as a function of screw speed, soybean flour concentration and MC, Li et al. (2005) showed that the highest ER can be achieved at 200 rpm, and that increasing screw speed to 275 and 350 rpm resulted in a drop in expansion. These results are in contrast to those of Seker (2005) who reported enhanced expansion with increasing screw speed.

N.J.E. de Mesa et al. / Journal of Food Engineering 90 (2009) 262–270

From the preceding discussion, it is clear that while studies have been done on the effect of formulation and process variables on expanded extrudates based on starch and soy protein, results have been conflicting and the underlying mechanism impacting the physico-chemical properties of these extrudates is unclear. Molecular level changes in starch and soy protein, and their interaction during extrusion are also not well understood. Hence, a systematic investigation on the interactions between starch and soy protein is needed to serve as a baseline. This would assist in maximizing soy protein levels in extrusion-puffed foods, either through starch modification or process improvement. The specific objective of this study was to evaluate the physical and microstructural characteristics of expanded extrudates based on native corn starch–soy protein concentrate blends, and relate these to the macromolecular changes in starch during extrusion. The latter was achieved using gel permeation chromatography (GPC) and standard measurements of water absorption and solubility indices. GPC helped in assessing the changes in starch molecular weight distribution brought about by extrusion.

263

average of three measurements was taken and BD was computed as in Eq. (2).

BD ¼

wsample : V container

ð2Þ

Piece densities (q) were obtained by dividing the mass of the sample (wpiece) by its volume (Vpiece). The latter was computed based on piece dimensions (diameter, d and length, l) measured using digital calipers (Digimatic Solar, Mitutoyo, Japan). Ten 2 cm long pieces were measured, and for each piece, the average of two diameter measurements was recorded. Eq. (3) shows the formula used.

q¼

wpiece 4wpiece ¼ V piece p d2 l

ð3Þ

Sectional expansion ratio (ER) was computed using Eq. (4). 2

ER ¼

d

2 ddie

ð4Þ

where ddie is the die diameter and d is the extrudate diameter. 2. Materials and methods 2.5. Image acquisition and void fraction 2.1. Materials Unmodified corn starch (CS) had 25% amylose and 75% amylopectin (Cargill Gel 03457, Cargill, Inc., Minneapolis, MN) and soy protein concentrate (SPC) had 71% protein, db (ProconTM 2000, The Solae Company, St. Louis, MO). CS–SPC blends were prepared using a batch ribbon blender (Wenger Manufacturing, Sabetha, KS), in the following ratios – 100:0, 95:5, 90:10, 85:15 and 80:20. 2.2. Extrusion system A Wenger TX-52 twin-screw extruder (Wenger Manufacturing, Sabetha, KS), with screw diameter of 52 mm, L/D ratio of 16:1, and a circular die with an opening of 4.75 mm was used to process all the blends. The extruder barrel had six heads, with the first five measuring 156 mm each in length and the last measuring 78 mm. The barrel temperatures and screw profile are shown in Fig. 1. The CS–SPC blends were extruded at two screw speeds, 230 and 330 rpm. The dry raw material feed rate was 75 kg/h. Water flow into the preconditioner and extruder were 5.0 kg/h and 4.7 kg/h, respectively, to achieve an in-barrel moisture of 22% (wb). Extruder conditions were allowed to stabilize for approximately 10 min before samples were collected. Ribbons of samples were collected from the extruder, dried at 55 °C for 3 h in a Thelco Laboratory Oven (Model 160 DM, Precision Scientific, Chicago, IL), and cut into 2.0 cm lengths. 2.3. Specific mechanical energy (SME) The mechanical energy input per unit mass of extrudate was calculated as follows

SME ¼

ðss0 Þ 100

 Prated  N N rated _ m

ð1Þ

where s is the measured torque, s0 is the no-load torque (assumed to be 0%), Prated is the rated power for the extruder (22.37 kJ/s), N is the measured extruder screw speed in rpm, Nrated is the rated extru_ is the mass flow rate (75 kg/h). der screw speed (336 rpm), and m 2.4. Measures of expansion Bulk density (BD) was obtained by taking the weight of the 2 cm samples (wsample) that filled a specified volume (Vcontainer = 1 L). The

Representative samples from each treatment were selected for image analysis. A desktop X-ray microtomography (XMT) imaging system (Model 1072, SkyScan, Aartselaar, Belgium) was used to scan the samples. The XMT was set at 20 kV/100 lA to obtain optimum contrast between solid and gaseous phases. For each sample, a set of three two-dimensional virtual ‘‘slices” were obtained after reconstruction. Calculations of cellular parameters were made using an image analysis software (Sigma Scan Pro, Systat, San Jose, CA). The total void area (Avoid) and cell wall area (Acell wall) were processed by the software, and the data were used to calculate void fraction (VF), as shown in Eq. (5). Details of XMT scanning, image reconstruction, thresholding, and microstructural parameters have been described previously (Trater et al. 2005).

%VF ¼

Avoid  100 Avoid þ Acell wall

ð5Þ

2.6. Mechanical properties Force-deformation data for each extrudate were obtained using a Texture Analyzer (Model TA-XT2i, Stable Micro Systems, Surrey, United Kingdom) fitted with a 25 kg load cell and a 38 mm diameter test probe. Using the compression mode, samples were compressed to 70% of their original height (20 mm) at a test speed of 10 mm/s. Thirty measurements were taken for each treatment. A force-deformation curve was obtained, and the number of peaks, n, integral of the curve, S (or area below the curve from 0% to 70% strain) and distance of compression, x, were computed. Using n, S and x values, average crushing force (Fcr) and crispness work (Wc) were calculated (Bouvier et al. 1997).

S F cr ¼ ðNÞ x S W c ¼ ðN mmÞ n

ð6Þ ð7Þ

2.7. Water absorption and solubility indices The procedure used by Gujral and Singh (2002) was used to determine the water absorption (WAI) and water solubility (WSI) indices. The extrudates were ground and sifted using a 250 lm sieve. Ground samples (wdry solid, 2.5 g) were dispersed in 25 mL

264

N.J.E. de Mesa et al. / Journal of Food Engineering 90 (2009) 262–270

Barrel Temperature (oC)

no control

30

30

80

100

120 Product discharge

Element No.

1a

2a

3

4

5

6

7

8

9 10 11 12

13

1=SEb-1-3/4-78; 2=SE-1-3/4-78; 3=SE-2-F-78; 4=SE-2-3/4-78; 5=SE-2-3/4-78; 6=KBc3-8.7-30; 7=SE-2-3/4-78; 8=KB-3-8.7-30; 9=SE-2-3/4-78; 10=KBc-3-8.7-30; 11=SE-21/2-52; 12 = SE-2-3/4-78; and 13=SE (conical)-2-3/4-78.d a

Left shaft elements are double flighted.

b SE = screw element Numbers: 1st – number of flights 2nd – relative pitch 3rd – element length, mm c KB = kneading block Numbers: 1st – number of elements 2nd – length of element, mm 3rd – angle of elements, degrees d

All screw and kneading block elements are forward and intermeshing.

Fig. 1. Screw and barrel temperature profiles of the TX-52 twin-screw extruder (Wenger Manufacturing, Sabetha, KS).

distilled water and continuously stirred for 30 min, then centrifuged at 3000 g for 10 min. The weight of the sediment was obtained and WAI was calculated. The amount of dried solids recovered after evaporating the supernatant in an oven at 135 °C for 2 h was used to determine WSI. Tests were done in triplicate and WAI and WSI were calculated as follows

wsedim ent  100 wdrysolid wdissolved solids in sup erna tan t WSI ¼  100 wdry solid

WAI ¼

ð8Þ ð9Þ

2.8. Gel permeation chromatography (GPC) The extrudates were ground and sifted using a 250 lm sieve. Ground extrudates (4.0 mg) were dispersed in 4.0 mL dimethyl sulfoxide (DMSO) and heated in a boiling-water bath for 24 h with continuous stirring. The solution was filtered through a 2.0 lm membrane, and then injected into an GPC system (PL-GPC 220, Polymer Laboratories Inc, Amherst, MA) equipped with differential refractive index (DRI) detector and phynogel 00H-0646-KO, 00H0644-KO, 00H-0642-KO columns (Phenomenex, Torrance, CA) connected in series. The columns and DRI detector were maintained at 80 °C. DMSO containing 5.0 mM NaNO3 was used as the mobile phase. Flow rate was maintained at 0.8 mL/min. Data were acquired and analyzed with a Cirrus GPC software version 3.0. A series of dextran standards (American Polymer Standards Corp., Mentor Ohio) representing molecular weights of 6.30  106, 4.00  106, 2.60  106, 846  103, 348  103, 131  103, 85.0  103, 36.7  103, 18.5  103, 3.40  103 were used for calibration. Duplicate tests were conducted for each treatment and representative curves were reported. 2.9. Experimental design and statistical analysis A 52 factorial experimental design, with five SPC (0%, 5%, 10%, 15% and 20%) and two extruder screw speeds (230 and 330 rpm) levels, was used for the production of expanded extrudates. For

every test conducted to characterize the extrudates, at least three replicate measurements were done per treatment, unless otherwise specified. The results were analyzed using analysis of variance (ANOVA) with the general linear model procedure (SAS version 9.1, SAS Institute, Cary, NC). When significant effects (p < 0.05) were indicated by ANOVA, Tukey pairwise comparisons were done to identify which treatments differed significantly (p < 0.05). Pearson’s coefficient of correlation (r) and their significances were determined for SME, ER, BD, q, Fcr, Wc and VF using the SAS Proc Corr procedure. The criteria defined by Franzblau (1958) was used to describe the degree of correlation (|r| < 0.20, negligible; |r| = 0.20–0.40, low; |r| = 0.40–0.60, moderate; |r| = 0.60–0.80, marked; |r| > 0.80, high).

3. Results and discussion 3.1. Specific mechanical energy SME is known to affect melt temperature and macromolecular degradation (Chinnaswamy and Hanna, 1990), and thus the expansion (Tang and Ding, 1994), and cellular and textural characteristics (Barrett and Peleg, 1992) of starch-based extrudates. In general, screw speed had a greater impact than SPC on SME (Fig. 2). SME at 330 rpm was higher than that at 230 rpm for all CS–SPC blends, except at 5% SPC. Increasing SME with increasing screw speed was attributed to higher shear rates, and is a well-documented observation in various studies involving extruded corn meal (Hsieh et al., 1990; Jin et al., 1994), corn starch (Agbisit et al., 2007), rice starch (Akdogan, 1996), rice flour (Lei et al., 2005) and protein–starch mixtures (Gropper et al., 2002). In general, the addition of 5–20% SPC to corn starch appeared to depress SME, although differences in SME at SPC levels greater than 10% were small. Up to 10% SPC, a drop in melt viscosity due to the fat and fiber contents of SPC most probably caused the reduction in SME. Beyond 10% SPC, protein interaction effects may have had an increased contribution, countering the effects of fat and fiber.

265

N.J.E. de Mesa et al. / Journal of Food Engineering 90 (2009) 262–270

350 330 rpm

SME (kJ/kg)

300

250

230 rpm

200

150 0

5

10

15

20

SPC Level (%) Fig. 2. Specific mechanical energy at varying levels of soy protein concentrate and screw speed.

3.2. Expansion and microstructural characteristics BD and ER data are shown in Figs. 3 and 4, respectively. In general, extrudate expansion was higher at 230 rpm than at 330 rpm, and increased with the addition of 5–20% SPC. At 0% and 5% SPC levels, BD’s at 230 and 330 rpm were comparable, but above 5% SPC, BD was significantly higher at the lower screw speed. BD increased as SPC level increased from 0% to 15%, with the exception of 10% SPC at 330 rpm. As SPC further increased to 20%, BD slightly decreased. The trends of q (data not shown) closely followed those of BD, and this was statistically supported by the Pearson’s correlation value of r = 0.98. ER trends were also similar to those for BD, and as expected, the two had a marked negative correlation (r = 0.77).

SME appeared to be the main factor that affected expansion of the extrudates. Both BD and q had a high negative correlation with SME (Table 1), ER also had a marked correlation with SME. Higher SME typically leads to increased melt temperature in the barrel, which in turn causes a greater driving force for expansion. Starch–SPC interactions also probably played an important role in affecting the expansion either indirectly through SME, or directly by disrupting the continuous starch matrix and thus reducing the extensibility of cell walls. In a study conducted on a whey protein–starch system, Allen et al. (2007) reported a reduction in expansion with increasing protein concentration and suggested that reduced expansion was a result of starch–protein interactions. Additionally, Zasypkin et al. (1992) found that ER increased significantly when starch was greater than 50% of the formulation in a

160

230 rpm

3

Bulk density (kg/m )

140

120

100

330 rpm

80

60 0

5

10

15

20

SPC Level (%) Fig. 3. Effect of varying levels of soy protein concentrate and screw speed on bulk density. *Error bars represent standard error of the mean of three measurements.

266

N.J.E. de Mesa et al. / Journal of Food Engineering 90 (2009) 262–270

20

Expansion ratio

15

10

330 rpm

230 rpm

5

0 0

5

10

15

20

SPC Level (%) Fig. 4. Effect of varying levels of soy protein concentrate and screw speed on expansion ratio. *Error bars represent standard error of the mean of 10 measurements.

Table 1 Pearson correlation coefficients, ra

ER

q BD Fcr Wc VF SME WAI a b

b

q

BD

Fcr

0.81*

0.77* 0.98*

0.98* 0.76* 0.70*

Wc

b

0.94* 0.73* 0.71* 0.96*

VF

SME

WAI

WSI

0.97* 0.79* 0.78* 0.97* 0.96*

0.66* 0.85* 0.86* 0.63 0.68 0.70*

0.67* 0.25 0.23 0.70 0.68 0.64* 0.09

0.73* 0.29 0.24 0.85* 0.77* 0.65* 0.06 0.94*

r values with asterisks represent significant correlations at p < 0.05. Fcr and Wc correlations are based on 0–15% SPC due to load cell limit at 20% SPC.

potato starch–soy protein isolate blend and inferred that protein restricted expansion. According to these authors, when the ratio of potato starch to soy protein isolate increases, starch forms a continuous matrix that enables water vapor to expand because starch melt viscosity is lower than protein melt viscosity. The work of Li et al. (2007) on phase inversion in CS–SPC mixtures also supports the supposition of having a more extensive protein network at higher levels of SPC. Similarly, the effects of phase inversion were described for potato starch and soybean protein isolate extrudates (Yuryev et al., 1995; Zasypkin et al., 1992). On the other hand, Chang et al. (2001) reported greater expansion with increasing SPC replacement of cassava starch. This is most probably because cassava starch, due to its lower amylose content, has a lower melt viscosity than corn starch at the same concentration. The X-ray micrographs (Fig. 5) show the underlying microstructural differences in the extrudates. It can be inferred visually that extrudate cells were larger in the case of higher screw speed. However, the cell size decreased and cell walls thickened with the addition of SPC up to 15%. Above this level, cell size slightly increased, possibly due to the dominance of protein–protein interactions. Interestingly, the number of cells (or cell density) increased with the addition of SPC. This was probably due to the foaming action of proteins, which led to greater number of nucleation sites for water vapor. An earlier study in our laboratory (Cheng et al., 2007) observed a similar increase in cell density and decrease in

cell size with the addition of whey protein isolate. While a concomitant thinning of cell walls was observed with the addition of whey protein isolate, the results of this study show that the opposite is true for SPC. This can be attributed to inherent differences in protein sources. X-ray microtomography allowed the calculation of VF, which is a relatively novel, microstructure-level measure of extrudate expansion. VF was calculated by taking into account the cell wall and void areas of two-dimensional XMT scans, and the data are shown in Fig. 6. Marked to high negative correlations existed between VF and the other more traditional measures of expansion (Table 1). The trends for VF with respect to screw speed and SPC level are similar to those for ER, BD and q. However, the overall effect of screw speed on VF was not significant, whereas the addition of SPC significantly affected the VF (p < 0.05). 3.3. Mechanical properties and their relationship to expansion characteristics Average crushing force (Fcr) data are shown in Fig. 7. At 0–5% SPC, (Fcr) at 230 and 330 rpm were very similar, but was higher for the lower screw speed beyond 5% SPC. However these differences were not significant. Regardless of screw speed, (Fcr) increased dramatically with the addition of 5% SPC and reached a plateau. No data was gathered beyond 15% SPC because the force required to crush the sample exceeded the load cell limit. Crispness work (Wc) (data not shown) showed the same trend as (Fcr), and the two were highly correlated (r = 0.96). Marked to high correlations were observed between Fcr or Wc and the various expansion parameters (BD, q, VF and ER). It is clear that decreased expansion led to higher Fcr and Wc due to the strengthening of the foam structure. This implies that more force and work were needed to deform the extrudates with shorter and thicker cell walls. This is consistent with the Gibson–Ashby theory of mechanical deformation of foams, as described by an earlier study on starch-based brittle foams in our laboratory (Agbisit et al., 2007). Other studies on expanded extrudates reported conflicting results for effect of soy protein on mechanical properties, because

267

N.J.E. de Mesa et al. / Journal of Food Engineering 90 (2009) 262–270

Fig. 5. X-ray microtomography cross-sectional images (same scale) of representative slices.

80 70

Void fraction (%)

60 50 330 rpm

40

230 rpm

30 20 10 0 0

5

10

15

20

SPC Level (%) Fig. 6. Effect of varying levels of soy protein concentrate and screw speed on void fraction. *Error bars represent standard error of the mean of three measurements.

of differences in expansion. For example, Li et al. (2005) reported increased extrudate hardness with the addition of soybean flour to corn flour. On the other hand, Chang et al. (2001) reported reduced extrudate hardness with increasing proportions of SPC in cassava starch–SPC blends due to higher expansion at these levels. 3.4. Starch macromolecular properties WSI and WAI are two important measures of physico-chemical changes in starch as a result of extrusion processing. WAI and WSI are related to the degree of starch fragmentation. Higher WAI indicates the presence of larger starch fragments, while higher WSI implies that starch has been dextrinized (Gomez and Aguilera 1984; Tang and Ding 1994; Seker, 2005; van den Einde et al., 2003). WAI and WSI data (Fig. 8) from the current study also showed this inverse relationship, as supported by the high negative correlation between the two parameters (Table 1). Generally, WSI was higher at 330 rpm, with the differences being significant at the two highest soy concentrations (Fig. 8).

Similarly, WAI was lower at higher screw speed, although significant differences were noted only at 10% and 20% SPC. These results suggest that higher screw speed degraded starch into smaller fragments, which are more soluble in water. The GPC results (Fig. 9a and b) confirmed these findings. All the curves of the extrudates at higher screw speed shifted to the right, indicating a decrease in starch molecular size distribution. The effect of screw speed on macromolecular degradation of starch is in agreement with the results obtained by Wen et al. (1990) for corn meal and van den Einde et al. (2004) for waxy corn starch. Although their work was conducted without protein addition, their conclusions are also valid in a binary system such as that used in this study. Even though increasing screw speed led to starch degradation, the effect of SPC was greater. In general, WAI decreased with increasing SPC levels, mainly because of a reduction in the starch content. Although WSI appeared to increase with increasing SPC, not all observed differences were significant. The GP chromatograms showed that the shift in molecular weight distribution was far greater without SPC than with SPC. For treatments contain-

268

N.J.E. de Mesa et al. / Journal of Food Engineering 90 (2009) 262–270

200 1 80

*

230 rpm 330 rpm

160

Crushing force (N)

140 1 20 100 80 60

*

40

values greater than those achieved with a load cell capacity of 25 k.g

20 0 0

5

10

15

20

SPC Level (%) Fig. 7. Effect of varying levels of soy protein concentrate and screw speed on crushing force. *Error bars represent standard error of the mean of 30 measurements.

10 WSI, 330 rpm

60

WSI, 230 rpm

50

9

40

7 6

30

5 20

WAI, 230 rpm

Water solubility index

Water absorption index

8

4 10

3

WAI, 330 rpm

2

0 0

5

10

15

20

SPC Level (%) Fig. 8. Effect of varying levels of soy protein concentrate and screw speed on water absorption and solubility indices. *Error bars represent standard error of the mean of three measurements.

ing SPC, although the chromatograms showed similar shifts towards smaller molecular sizes at higher screw speed, the curves at 5% SPC were almost identical and the differences were only evident at higher SPC levels. The small impact of screw speed on starch fragmentation partly echoes the results of Seker (2005), who reported that screw speed did not significantly affect the WAI and WSI of starch–soy protein isolate mixtures. While Chang et al. (2001) reported higher WAI and lower WSI with increasing levels of SPC in cassava starch–SPC blends, there was no reason provided to explain their results.

4. Conclusions Screw speed and SPC levels affected the expansion, mechanical, and macromolecular properties of CS–SPC extrudates. Expansion was primarily affected by SME. Increasing SME led to higher melt temperature and greater steam flash off and expansion at the end of the die. XMT images showed the underlying microstructure of extrudates. The addition of SPC resulted in a reduction in cell size yet with in an increase in the number of cells, which was

N.J.E. de Mesa et al. / Journal of Food Engineering 90 (2009) 262–270

269

Fig. 9a. Molecular weight distributions of native and extruded corn starch.

Fig. 9b. Molecular weight distributions of corn starch–soy protein concentrate (80:20) extruded at 230 rpm and 330 rpm.

attributed to the foaming action of proteins. Mechanical strength had an inverse relationship with expansion, and increased with the addition of SPC probably due to reduced expansion (resulting from lower SME) and cell wall thickening. Extrudates with larger voids and thinner cell walls had lower Fcr and Wc due to structural weakening. The effect of increasing SPC content was greater than the effect of screw speed on starch degradation (WAI, WSI and starch molecular weight shifts). The role of starch and protein in compounded formulations for expanded snacks and breakfast cereal can be more effectively explained by studying the combination of simple systems of starch and protein, as was done in this study. Knowledge of the consequences of processing conditions and interactions between native starch and soy protein in extrusion-expanded snacks will serve as a reference for future studies on soy protein – fortified extrudates using various starches (e.g., highamylose starch) and process conditions. Acknowledgements The authors would like to thank Mr. Eric Maichel for his technical assistance and Wenger Manufacturing, Inc. for their continued support to the extrusion laboratory at Kansas State University. This is Contribution Number 08-197-J from the Kansas Agricultural Experiment Station, Manhattan, KS. References Agbisit, R., Alavi, S., Cheng, E., Herald, T., Trater, A., 2007. Relationships between microstructure and mechanical properties of cellular cornstarch extrudates. Journal of Texture Studies 38, 199–219.

Akdogan, H., 1996. Pressure, torque and energy responses of a twin-screw extruder at high moisture contents. Food Research International 29 (5–6), 423–429. Allen, K.E., Carpenter, C.E., Walsh, M.K., 2007. Influence of protein level and starch type on an extrusion-expanded whey product. International Journal of Food Science and Technology 42, 953–960. Barrett, A.H., Peleg, M., 1992. Extrudate cell structure–texture relationships. Journal of Food Science 57 (5), 1253–1257. Bouvier, J.M., Bonneville, R., Goullieux, A., 1997. Instrumental methods for the measurement of extrudate crispness. Agro-food-Industry Hi-Tech 8 (1), 16–17. Chang, Y.K., Hashimoto, J.M., Acioli-Moura, R., Martinez-Flores, H.E., MartinezBustos, F., 2001. Influence of extrusion conditions on cassava starch and soybean protein concentrate blends. Acta Alimentaria 30 (2), 189–203. Cheng, E.M., Alavi, S., Pearson, T., Agbisit, R., 2007. Mechanical-acoustic and sensory evaluations of cornstarch-whey protein isolate extrudates. Journal of Texture Studies 38, 473–498. Chinnaswamy, R., Hanna, M.A., 1990. Macromolecular and functional properties of native and extrusion-cooked corn starch. Cereal Chemistry 67 (5), 490–499. e-CFR 10182 1999. Health claims: soy protein and risk of coronary heart disease (CHD). 64 FR 57732, October 26, 1999. Electronic-Code of Federal Regulations, Title 21: Food and Drugs, 101.82. http://ecfr.gpoaccess.gov. Online access: 19 November 2007. Faubion, J.M., Hosney, R.C., 1982. High temperature short-time extrusion cooking of what starch and flour, II: Effect of protein and lipid on extrudate properties. Cereal Chemistry 59 (6), 533–537. Franzblau, A., 1958. A Primer of Statistics for Non-statisticians. Harcourt, Brace and World, New York. Friedman, M., Brandon, D.L., 2001. Nutritional and health benefits of soy proteins. Journal of Agricultural and Food Chemistry 49 (3), 1069–1086. Ghorpade, V.M., Bhatnagar, S., Hanna, M.A., 1997. Structural characteristics of corn starches extruded with soy protein isolate or wheat gluten. Plants for Human Nutrition 51, 109–123. Gomez, M.H., Aguilera, J.M., 1984. A physicochemical model for extrusion of corn starch. Journal of Food Science 49, 40–43. 63. Gropper, M., Moraru, C.I., Kokini, J.L., 2002. Effect of specific mechanical energy on properties of extruded protein–starch mixtures. Cereal Chemistry 79 (3), 429– 433.

270

N.J.E. de Mesa et al. / Journal of Food Engineering 90 (2009) 262–270

Gujral, H.S., Singh, N., 2002. Extrusion behaviour and product characteristics of brown and milled rice grits. International Journal of Food Properties 5 (2), 307– 316. Hsieh, F., Peng, I.C., Huff, H.E., 1990. Effects of salt, sugar and screw speed on processing and product variables of corn meal extruded with a twin screw extruder. Journal of Food Science 55, 224–227. Jin, Z., Hsieh, F., Huff, H.E., 1994. Extrusion cooking of corn meal with soy fiber, salt, and sugar. Cereal Chemistry 71 (3), 227–234. Lei, H., Fulcher, R.G., Ruan, R., van Lengerich, B., 2005. SME-Arrhenius model for WSI of rice flour in a twin-screw extruder. Cereal Chemistry 82 (5), 574–581. Li, J.-Y., Yeh, A.-I., Fan, K.-L., 2007. Gelation characteristics and morphology of corn starch/soy protein concentrate composites during heating. Journal of Food Engineering 78, 1240–1247. Li, S.-Q., Zhang, H.Q., Jin, Z.T., Hsieh, F.-H., 2005. Textural modification of soya bean/ corn extrudates as affected by moisture content, screw speed and soya bean concentration. International Journal of Food Science and Technology 40, 731– 741. Liu, K., 1997. Soybeans: Chemistry, Technology and Utilization. Chapman & Hall, New York, NY. Messina, M., Barnes, S., 1991. The role of soy products in reducing risk of cancer. Journal of the National Cancer Institute 85, 541–546. Riaz, M.N., 2006. Soy Applications in Food. Taylor and Francis Group, LLC, Boca Raton, FL. p. 70. Seker, M., 2005. Selected properties of native or modified maize starch/soy protein mixtures extruded at varying screw speed. Journal of the Science of Food and Agriculture 85, 1161–1165.

Tang, J., Ding, X.L., 1994. Relationship between functional properties and macromolecular modification of extruded corn starch. Cereal Chemistry 71 (4), 364–369. Trater, A.M., Alavi, S., Rizvi, S.S.H., 2005. Use of non-invasive X-ray microtomography for characterizing microstructure of extruded biopolymer foams. Food Research International 38, 709–719. van den Einde, R.M., Bolsius, A., van Soest, J.J.G., Janssen, L.P.B.M., van der Goot, A.J., Boom, R.M., 2004. The effect of thermomechanical treatment on starch breakdown and the consequences for process design. Carbohydrate Polymers 55, 57–63. van den Einde, R.M., van der Goot, A.J., Boom, R.M., 2003. Understanding molecular weight reduction of starch during heating–shearing processes. Journal of Food Science 68 (8), 2396–2404. Wen, L.-F., Rodis, P., Wasserman, B.P., 1990. Starch fragmentation and protein insolubilization during twin-screw extrusion of corn meal. Cereal Chemistry 67 (3), 268–275. Yuryev, V.P., Zasypkin, D.V., Alexeev, V.V., Bogatyryev, A.N., 1995. Expansion ratio of extrudates prepared from potato starch–soybean protein mixtures. Carbohydrate Polymers 26, 215–218. Zasypkin, D.V., Lee, T.-C., 1998. Extrusion of soybean and wheat flour as affected by moisture content. Journal of Food Science 63 (6), 1058–1061. Zasypkin, D.V., Yuryev, V.P., Alexeyev, V.V., Tolstoguzov, V.B., 1992. Mechanical properties of the products obtained by the thermoplastic extrusion of potato starch–soybean protein mixtures. Carbohydrate Polymers 18, 119–124.