Pasta-Like Product from Pea Flour by Twin-Screw Extrusion

Transcription

1 JOURNAL OF FOOD SCIENCE Pasta-Like Product from Pea Flour by Twin-Screw Extrusion N. Wang, P.R. Bhirud, F.W. Sosulski and R.T. Tyler ABSTRACT The effects of extrusion parameters on the characteristics of a pasta-like product were investigated. Increasing moisture content of the dough decreased brightness, bulk density, cooking loss and stickiness, but increased cooking time and firmness. Raising the barrel temperature increased cooking time, firmness and stickiness, but brightness, bulk density and cooking loss decreased. Increasing the screw speed increased cooking time, cooking loss, firmness and stickiness, but decreased brightness and bulk density. The product exhibited superior integrity, flavor, and texture after cooking, and less change after overcooking, compared to pasta-like products prepared from pea flour using a conventional pasta extruder. Key Words: extrusion cooking, pea pasta, gluten, starch gelatinization, microstructure INTRODUCTION PASTA PRODUCTS ARE TRADITIONALLY manufactured by blending durum wheat semolina and water to form an homogeneous mixture. This is kneaded to form a dough, which is extruded into the desired shape at room temperature and atmospheric pressure or under vacuum, and subsequently dried (Aktan and Khan, 1992). Some humans cannot tolerate wheat gluten, which has been linked to celiac disease, a specific disorder of intestinal absorption (Kowlessar, 1972). The prescribed treatment is strict adherence to a gluten-free diet (Rosenlund, 1970). Pasta-like products made solely from rice flour or corn flour in lieu of wheat flour or semolina do not adequately simulate the chewability of durum-based products, and have poor cooking quality. The use of field pea flour in the manufacture of such pasta-like products would be advantageous, since it is higher in protein and lysine than wheat flour or semolina, and gluten-free. In our laboratory, preparation of a pastalike product from pea flour using a conventional pasta extruder has been tested. The product had poor quality, particularly in texture, low acceptability of flavor, and sensory properties, and notable disintegration during cooking. A process is needed for the manufacture, from pea flour, of pasta-like products with improved texture, flavor, and cooking quality. Authors Wang, Bhirud and Tyler are affiliated with the Dept. of Applied Microbiology & Food Science, Univ. of Saskatchewan, 51 Campus Drive, Saskatoon, SK, Canada S7N 5A8. Author Sosulski is with the Dept. of Plant Sciences, Univ. of Saskatchewan, 51 Campus Drive, Saskatoon, SK, Canada S7N 5A8. Direct inquiries to Dr. N. Wang. Extrusion technology has led to production of a wide variety of cereal-based foods, including snacks and ready-to-eat breakfast cereals (Bailey et al., 1991). Extrusion cooking is a high temperature, short time (HTST) process, whereby starches are gelatinized, proteins are denatured, and extrudates are texturally and histologically restructured (Smith, 1971). HTST extrusion reduces microbial contamination and inactivates enzymes (Fellows, 1988; Likimani et al., 1990). Substantial efforts have been reported to develop wheat-based quick cooking pasta products by extrusion cooking (Lomttillo and Wolcott, 1983; Wenger and Huber, 1991). However, none of the reported methods provided a way of making pasta products from legume flours. Our objectives were: (1) to investigate the effects of extrusion processing variables, such as dough moisture, barrel temperature, screw speed and feed rate, on pasta-like product characteristics, and (2) to evaluate the cooking quality of such pasta-like products made entirely from pea flour by twin-screw extrusion. MATERIALS & METHODS Materials Starch-enriched, air-classified yellow pea flour from Parrheim Foods, (Saskatoon, Saskatchewan, Canada), was used. Procedures for dehulling, pin milling and air classifying yellow pea have been described by Tyler et al. (1981). The air-classified pea flour contained 9.2% moisture and, on a moisturefree basis, 12.3% protein, 69.1% starch and 4.8% dietary fiber. Two commercial spaghetti products, one fresh and one dry, were obtained from local food stores for comparison. The fresh pasta product was dried at room temperature prior to evaluation. Extrusion A twin-screw, co-rotating extruder (Model ZSK 57, Werner & Pfleiderer, Ramsey, NJ) was used. Specifications were: screw dia, 57 mm; length-dia ratio, 26:1; die openings, circular and either 2.38 mm or 1.09 mm dia (Table 1). The screws were operated at rpm. The extruder barrel consisted of 8 segmented zones. The first segment provided an inlet for the mixture of raw materials from the feeder. Barrel temperature was maintained at 30 C in the first segment, at 90 C in the second and third segments (T 1 ), and at 95 C in the seventh and eighth segments (T 3 ). Temperatures in the fourth through sixth segments (T 2 ) were adjusted to 90, 100 or 110 C. Pea flour was fed into the extruder at kg/h using a K-tron T-35 twin-screw volumetric feeder (K-tron Corp., Glassboro, NJ). Water was fed into the extruder at 4.45 to 7.85 kg/h, so that the final mixture moisture content was 24.3 to 31.8%. The pasta-like (spaghetti-type) products, containing 18 to 23% moisture, were cut to length ( 30 cm) with a knife as they exited the extruder die and dried at ambient temperature to a final moisture of 10%. Table 1 Screw configuration for preparation of a pasta-like product from pea flour by twin-screw extrusion Element type No. of elements 30/30 a 1 80/ / /80 4 KB 45 /5/40 b 1 80/ /20 LH c 1 60/60 1 KB 45 /5/ / /20 LH 1 30/ /20 LH 1 30/ /20 LH 1 30/ / / /40 3 ascrew elements: pitch (mm)/length (mm). bkneading blocks (KB): stagger /number of disks/length (mm). c LH = left-handed Institute of Food Technologists Volume 64, No. 4, 1999 JOURNAL OF FOOD SCIENCE 671

2 Pasta-Like Product From Pea Flour... Table 2 Effects of extrusion variables on some properties of pasta-like products from pea flour by twin-screw extrusion Experiments b Process variables a M (%) T ( C) N (rpm) F (kg/h) Responses Physical properties L a b Expansion ratio Bulk density (g/cm 3 ) Nutritional properties DSG (%) TIA reduction (%) Cooking properties Cooking time (min) Cooked weight (g) Cooking loss (%) Textural properties Firmness (g cm) Stickiness (N/m 2 ) Compressibility (%) Recovery (%) am = moisture content (%); T = barrel temperature ( C); N = screw speed (rpm); F = feed rate (kg/h); L = coefficient dark (0) to bright (+); a = coefficient green (-) to red (+); b = coefficient blue (-) to yellow (+); DSG = degree of starch gelatinization (%); TIA = trypsin inhibitor activity (mg/g). bcircular die opening of 2.38 mm dia. Experimental design A fractional factorial design (Myers, 1971) was employed to investigate the influence of moisture content (M, %), barrel temperature (T, C), screw speed (N, rpm) and feed rate (F, kg/h). The experimental design generators were F M T N, wherein the alias structure was I1M T N F, M T N F, T M N F, N M T F, F M T N, M T N F, M N T F, M F T N. A total of 11 experiments (Table 2) were conducted with 3 replicates at center levels. The following model (Eq (1)) was fitted to the experimental data and analyzed with Minitab statistical analysis software (Minitab, 1995): Y i b o b M M b T T b N N b F F b M T M T b M N M N b M F M F (1) where Y i is the response (physical, nutritional, cooking or textural properties), b o is a constant, and b M, b T, b N, b F, b M T, b M N and b M F are regression coefficients. For each response, 3-dimensional plots were produced from the equation by holding two variables constant and varying the other two. Physical tests on pasta-like products Measurement of expansion ratio and color. Expansion ratio was calculated as the ratio of the diameter of the pasta-like (spaghetti-type) product (average of 6 measurements) to the die diameter. The diameter of the product was measured using a caliper. Unit bulk density was determined from the weight and volume of 1.0-cm-long pieces of product (average of 6 pieces). For color measurements, samples were broken into 1.5-cm-long pieces and placed in the sample cup, care being taken to ensure the bottom surface was fully covered. Pasta color was measured on a HunterLab Color Difference Meter (ColorQUEST, Hunter Associates Laboratory, Inc., Reston VA) as L, a and b values standardized with a white color standard, where L is the brightness coefficient [dark (0) to bright ( )], a the coefficient from green (-) to red ( ), and b the coefficient from blue (-) to yellow ( ). Illuminant D65 was used. Quality evaluation of cooked pastalike product Cooking time, cooked weight and cooking loss. Cooking tests were performed on duplicate samples according to Matsuo et al. (1992). Product (5g dry wt) was added to 125 ml of rapidly boiling deionized water. The product was deemed cooked when the white core in a strand, when pressed between two glass slides, was no longer visible. The time required to reach this point for each sample was noted as cooking time. Cooked weight was determined on 5 g (initial weight) of cooked product that had been cooked at its previously determined cooking time and drained on a sieve for 5 min. Cooking loss was expressed as the ratio of solids lost in the cooking water to the dry weight of initial product. Solids in the cooking water were recovered by freeze drying. Product firmness. Cooked product firmness was measured according to Walsh (1971) with a TA-XT2 texture analyzer (Texture Technologies Corp., Scarsdale, NY) interfaced with a personal computer running SMS1 software (Stable Micro Systems, Haslemere, UK). The load cell was of 25 kg capacity and 1.0g sensitivity. Each product (10g) was broken into strands 5 cm in length, cooked for its determined cooking time in 300 ml of deionized water, and drained for 1 min on a sieve. Single strands of cooked product were placed on a sample holder and sheared at a 90 angle with a specially designed polycarbonate resin cutting tooth (Walsh, 1971). A continuous recording of force vs distance was plotted on the computer screen (Fig. 1A). The area under the force vs distance curve (g-cm), equivalent to the amount of work required to shear the strand of cooked product, was noted as the firmness value. Values reported are averages of triplicate determinations. Product stickiness. Cooked product stickiness was determined by a modification of the method of Dexter et al. (1983) with the TA-XT2 texture analyzer. Product (10g) was broken into strands 5 cm in length and cooked for its determined cooking time in 300 ml of deionized water. The cooked product was drained for 1 min on a sieve and loaded onto a polished aluminum plate. Strands were laid side-by-side to cover an area of 5 cm in width. Stickiness measurements were performed within 10 min of loading. Three minutes before testing, an aluminum cover plate was placed over the strands, and excess water was blotted with tissue paper. The cover plate was also made of polished aluminum of equivalent dimensions, but had a rectangular opening in its center to allow plungerto-sample contact. The plunger, also rectangular, had a polished aluminum contact surface, and was attached to a load cell of 25 kg capacity and 1.0g sensitivity. System parameters were: (a) arm speed prior to sample contact, 2.0 mm/sec; (b) trigger point, 5.0 g- force; (c) arm speed during compression of strands, 0.1 mm/sec; (d) compression force at the end of compression, 400 g-force; (e) holding time at maximum compression, JOURNAL OF FOOD SCIENCE Volume 64, No. 4, 1999

3 sec; (f) arm speed during retraction and sensing of stickiness, 3.0 mm/sec. The force applied during compression of the cooked pasta, and on lifting the plunger, was plotted (Fig. 1B). Stickiness was defined as the maximum depression recorded during lifting of the plunger. The stickiness value was converted to N/m 2 using a value of 5,200 N/m 2 for a compression force of 400g. Values reported are averages of duplicates. Compressibility and recovery of cooked product. Compressibility and recovery of cooked product were measured according to Matsuo and Irvine (1971) with the TA-XT2 texture analyzer. A single cooked product strand, about 2-3 cm in length, was placed in the slot of the sample holder. The strand was compressed to a stress of N/m 2 for 15 sec. The compression force was removed and the compression-recovery curve was recorded (Fig. 1C). Values are averages of duplicates. Compressibility was defined as the ratio of Y to X, and recovery as the ratio of the distance the blade was forced back (Y-Z) to the penetration, Y. X represents the diameter of the cooked product strand (mm), Y the distance of penetration of the blade into the sample (mm), and Z the extent to which the elastic component forced the blade back (mm). Chemical analysis Trypsin inhibitor activity (TIA) levels were determined by the method of Smith et al. (1980). Degree of starch gelatinization (DSG) was determined by the method of Chiang and Johnson (1977). Protein (N 6.25), moisture and starch were determined by AACC methods 46-11A, 44-15A and 76-13, respectively (AACC, 1995). Total dietary fiber was determined by the AOAC (1990) method. Microscopy Cross and longitudinal sections of the pasta-like product from pea flour by twinscrew extrusion and of commercial spaghetti were examined by scanning electron microscopy (SEM). Extrusion conditions used in preparation of the pasta-like product were: dough moisture content, 32.0%; barrel temperature, 110 C; screw speed, 125 rpm; circular die opening, 1.09 mm dia. Two specimens of each sample were examined. Samples were mounted on aluminum studs with cross and longitudinal sections exposed. The exposed surface of one of the two specimens was wetted with a drop of water. After 3-5 min, the excess water on the surface was blotted away with tissue paper. The wet specimens were frozen in liquid nitrogen and freeze-dried. All specimens were ultimately coated with gold using a sputter coater (S150B, Edwards, Wilmington, MA) at 1.0 kv and 40 ma for 3 min, prior to examination by SEM (SEM 505, Philips, Eindhoven, Holland) at 30 kv. RESULTS & DISCUSSION REGRESSION COEFFICIENTS WERE DETERmined from fitting the experimental data (Table 2) to Eq (1) (Table 3). Analysis of variance indicated that the model was acceptable Fig. 1 Typical plots for determination of firmness [A], stickiness [B] and compressibility and recovery [C] of cooked pasta-like products from pea flour. X=dia of cooked product (mm), Y=distance of penetration of the blade into the sample (mm), Z=the extent to which the elastic component forces the blade back (mm). Fig. 2 Effects of processing variables on L, a and b color values of pasta-like products from pea flour. Volume 64, No. 4, 1999 JOURNAL OF FOOD SCIENCE 673

4 Pasta-Like Product From Pea Flour... Table 3 Regression coefficients and R 2 values for some properties of pasta-like products from pea flour by twin-screw extrusion Regression coefficients b Properties a b o b M b T b N b F b M T b M N b M F R 2 Physical properties L ** -2.69** ** * * a ** * * ** ** b ** * * E pansion ratio -4.16** ** * ** ** Bulk density (g/cm 3 ) 2.75** * * ** * Nutritional properties DSG (%) ** -3.61** 1.32* -2.66** * ** ** TIA reduction (%) ** 3.44* 1.65** * * Cooking properties Cooking time (min) ** 9.30** * 1.95** ** * Cooked weight (g) 9.63** * ** Cooking loss (%) ** * ** * Te tural properties Firmness (g cm) ** * * * * Stickiness (N/m 2 ) ** ** * -2.63* * Compressibility (%) ** ** -4.38* -1.01** ** Recovery (%) ** 7.69** 4.73* -1.11* ** * * a L = coefficient dark (0) to bright (+); a=coefficient green (-) to red (+); b=coefficient blue (-) to yellow (+); DSG=degree of starch gelatinization (%); TIA=trypsin inhibitor activity (mg/g). b M = moisture content (%); T = barrel temperature ( C); N = screw speed (rpm); F = feed rate (kg/h); b o = constant; b M, b T, b N, b F, b M T, b M N and b M F = regression coefficients. ***, **, * = significant at p<0.001, p<0.01, and p<0.05, respectively. +++Blank spaces indicate coefficient does not contribute significantly to the model. (p 0.05) and could be used to predict values for the 14 response variables derived during the manufacture of the pasta-like products by high temperature extrusion. Effect of variables on properties of pasta-like products The influence of processing variables was determined on brightness, redness and yellowness of pasta-like products prepared from pea flour by high temperature extrusion (Fig. 2). Brightness (L value) decreased as moisture content, barrel temperature or screw speed was increased (Fig. 2A and 2B). Redness increased as moisture content increased, but decreased as barrel temperature or screw speed increased (Fig. 2C and 2D). Yellowness decreased as barrel temperature or screw speed increased (Fig. 2E and 2F). Moisture content had little effect on yellowness. The development of color in the pasta-like product would have been due, in part, to Maillard-type reactions during high temperature extrusion. Moisture and temperature conditions would have been favorable (Berset, 1989), and peas contain substantial levels of soluble carbohydrates and lysine. This is the most reactive amino acid with respect to the Maillard reaction. Feed rate had no effect on brightness, redness or yellowness (Table 3). Effects of extrusion processing variables were determined (Fig. 3) on the expansion ratio, bulk density, degree of starch gelatinization (DSG) and reduction of trypsin inhibitor activity (TIA). The expansion ratio increased with dough moisture content (Fig. 3A and 3B). Increased die pressure was observed (data not shown) as moisture content Fig. 3 Effects of processing variables on expansion ratio, bulk density, degree of starch gelatinization (DSG) and reduction of trypsin inhibitor activity (TIA) of pasta-like products from pea flour. 674 JOURNAL OF FOOD SCIENCE Volume 64, No. 4, 1999

5 increased, which would increase the degree of expansion. The increased die pressure may have resulted from an increase in moisture which caused a reduction in mass temperature within the extruder barrel and, as a result, an increase in dough viscosity. Expansion ratio increased as barrel temperature and screw speed increased (Fig. 3A and 3B). The increase in barrel temperature would create a higher vapor pressure in the dough, resulting in more flashing of moisture as the product exited the die and, consequently, greater expansion. Increasing screw speed would increase shear rate and, therefore, extrudate temperature and the degree of expansion. Bulk density decreased as moisture content, barrel temperature and screw speed increased (Fig. 3C and 3D), which was expected due to the inverse relationship between expansion ratio and bulk density. The degree of starch gelatinization (DSG) is determined by moisture, temperature and shear during extrusion-cooking. As expected, an increase in moisture content increased DSG (Fig. 3E and 3F). Increasing the moisture in the dough would increase both granule swelling and amylose leaching and, as a result, DSG. The degree of starch gelatinization also increased as barrel temperature increased (Fig. 3E). Starch gelatinization has been shown to follow first-order kinetics and was dependent on temperature (Lund, 1989). The effect of screw speed on DSG was complex (Fig. 3F). At lower moisture contents, DSG increased with decreasing screw speed which would increase residence time. At higher moisture, DSG increased with increasing screw speed which would reduce residence time but enhance the rate of shear. The resultant higher mechanical and thermal energy input would increase DSG. The extent of reduction in TIA increased as barrel temperature or screw speed was increased, but decreased when moisture content increased (Fig. 3G and 3H). Similar results had been reported by Aguilera and Kosikowski (1976) in texturized soy products, and by Wang et al. (1999) in texturized air-classified pea protein. Feed rate had little effect on expansion ratio, bulk density, DSG or TIA reduction. The effects of extrusion variables on cooking time, cooked weight and cooking loss were compared (Fig. 4). Cooking time increased with increasing dough moisture, barrel temperature or screw speed (Fig. 4A and 4B). A strong positive correlation (r 0.909, P<0.01) existed between expansion ratio and cooking time. Cooked weight (water absorption) increased when dough moisture increased, but decreased as screw speed increased (Fig. 4C and 4D). Anderson et al. (1969) reported that the water absorption of extruded cereal products was higher when prepared from higher moisture doughs. Diosady et al. (1985) reported that the lower water absorption observed for extruded starch produced at higher screw speed might be related to increased shear, resulting in structural modification of the starch. Barrel temperature and feed rate had little effect on cooked weight. Cooking loss decreased as dough moisture content or barrel temperature increased, but increased as screw speed increased (Fig. 4E and 4F). Feed rate had no effect on cooking loss. It is generally accepted that extrusion cooking of starch or starchy materials involves extensive degradation of macromolecules (Colonna and Mercier, 1983). The macromolecular degradation of both amylose and amylopectin, leading to lower molecular weight material, would result in increased water solubility of the extrudate. The degradation of starch is greater at lower dough moistures and higher screw speeds (Colonna and Mercier, 1983). The influences of extrusion variables on textural properties of the cooked pasta-like products prepared from pea flour were determined (Fig. 5). Firmness increased when dough moisture, barrel temperature or screw speed increased (Fig. 5A and 5B). Firmness of the pasta-like product correlated positively (r 0.839, P<0.01) with its expansion ratio, which was greater at higher dough moistures, barrel temperatures and screw speeds (Fig. 3A and 3B). Stickiness of the cooked pasta-like product decreased as moisture content increased, but increased when either barrel temperature or screw speed increased (Fig. 5C and 5D). Dexter et al. (1985) indicated amylose on the surface of cooked spaghetti may be a contributing factor to surface stickiness. The effects of feed rate on firmness and stickiness were not significant (Table 3). The compressibility of the cooked pasta-like product decreased, but recovery increased, as dough moisture, barrel temperature or screw speed increased (Fig. 5E-5H). Compressibility correlated negatively (r , P<0.05) with expansion ratio. Feed rate had no effect on compressibility or recovery (Table 3). Fig. 4 Effects of processing variables on cooking time, cooked weight and cooking loss of pasta-like products from pea flour. Volume 64, No. 4, 1999 JOURNAL OF FOOD SCIENCE 675

6 Pasta-like Product From Pea Flour Table 4 Some properties of pasta-like products from pea flour compared with commercial wheat pasta a Pasta-like product Commercial wheat pasta e Properties b Low temp c High temp d Fresh Dry Physical properties L a d b c a 4.17 d a 5.35 c 5.72 b b d c b a Cooking properties Cooking time (min) 17 a 10 c 13 b 13 b Cooked weight (g) 17.4 a 14.7 c 13.6 d 15.2 b Cooking loss (%) 48.2 a 20.5 b 7.1 c 7.8 c Te tural properties Firmness (g cm) N/D f 15.8 a 14.7 a 8.8 b Stickiness (N/m 2 ) N/D b c a Compressibility (%) N/D 74.6 c 100 a 81.3 b Recovery (%) N/D 60.4 a 0.0 c 42.1 b ameans in rows with same letter not significantly different (p>0.05) using Duncan s multiple range test. bl=coefficient dark (0) to bright (+); a=coefficient green (-) to red (+); b=coefficient blue (-) to yellow (+). clow Temp=product from pea flour with conventional low temperature pasta extruder at moisture 30.0% and at room temperature; circular die opening, 1.2 mm dia. d High Temp=product from pea flour by twin-screw extrusion at moisture, 32%; barrel temperature, 110 C; screw speed, 125 rpm; circular die opening, 1.09 mm dia. The pasta-like product had 1.52 mm dia. efresh=commercial fresh spaghetti with 1.48 mm dia; Dry=commercial dry spaghetti with 1.50 mm dia. fn/d=not determined due to disintegration during cooking. Fig. 5 Effects of processing variables on firmness, stickiness, compressibility and recovery of pasta-like products from pea flour. Comparison of pasta-like products from pea flour by extrusion with commercial wheat pasta The characteristics of pasta-like (spaghetti-type) products from pea flour by twinscrew extrusion or by low temperature extrusion were compared with those of commercial spaghetti products (Table 4). The pasta-like product from twin-screw extrusion was less bright, more red and more yellow than that from conventional low temperature extrusion. When cooked, the twin screw product exhibited a shorter cooking time, lower cooked weight, lower cooking loss and more desirable texture. Informal sensory tests (data not shown) indicated that the pasta-like product from twin-screw extrusion had more acceptable flavor after cooking than that prepared by conventional low temperature extrusion. The commercial spaghetti products were brighter than the pasta-like product from pea flour by twin-screw extrusion, but the pea flour product was more red and more yellow. The cooking time of the pea flour product was shorter than those of the commercial spaghetti products, whereas the cooked weight was within the range exhibited by the commercial products. However, the pastalike product from pea flour had a much greater cooking loss than did the commercial spaghetti products, due to the absence of a gluten matrix. The cooked pasta-like product was as firm as, and stickier than, the cooked fresh spaghetti, but firmer and much less sticky than the cooked dry commercial spaghetti. The pasta-like product from pea flour was less compressible than the commercial spaghetti products, but had a higher recovery of volume. Microstructure of pasta-like product from pea flour The microstructure of cross and longitudinal sections of a commercial spaghetti product was compared with the pasta-like product from pea flour by extrusion cooking (Fig. 6). The surface of the commercial spaghetti appeared to be dense and compact, and coated with a smooth protein film with few openings (Fig. 6a and 6b). Some starch granules of varying size were visible. Many cracks and small holes were apparent in the protein matrix, which would permit rapid penetration of cooking water. The pasta-like product (Fig. 6c and 6d) also exhibited a dense, compact structure, with relatively few swollen starch granules. These appeared to be completely coated with a gelatinized starch and protein matrix. After the surfaces had been treated with water, differences in microstructure between the two products became more apparent (Fig. 7). The cross section of the commercial spaghetti product revealed a marked change in internal structure after treatment with water (Fig. 7a). The interior was altered from a compact structure with few visible starch 676 JOURNAL OF FOOD SCIENCE Volume 64, No. 4, 1999

7 granules (Fig. 6a) to a more porous structure with starch granules loosely held within a discontinuous protein matrix (Fig. 7a). Starch granules were readily apparent and the protein matrix had retracted into a discontinuous phase, although most of the starch granules appeared to be held quite tightly. In con- trast, the product from pea flour revealed, after treatment with water, a compact structure with swollen starch granules imbedded in a gelatinized starch and protein matrix (Fig. 7c). A longitudinal section of the extrusion-cooked product (Fig. 7d) clearly showed that the swollen starch granules were surrounded by a laminar gelatinized starch and protein matrix, and seemed to be aligned in the direction of flow through the extruder barrel. The starch granules of the commercial spaghetti product were imbedded in a gluten network (Fig. 7b). CONCLUSIONS PASTA-LIKE PRODUCTS WERE PREPARED from pea flour by high temperature extrusion. Variables such as dough moisture, barrel temperature and screw speed had notable effects on physical, textural and cooking characteristics of the pasta-like products. Compared to commercial spaghetti, the pea flour products had shorter cooking times, were firmer and less sticky, and had higher cooking losses. Extrusion-cooked pasta-like pea flour products had better integrity, more acceptable flavor, and superior texture after cooking, and less change after overcooking, than those prepared using a conventional low temperature pasta extruder. The extrusioncooked pea flour product had a compact structure with relatively few swollen starch granules deeply imbedded in a gelatinized starch and protein matrix and aligned in the direction of flow through the extruder barrel. Products from pea flour by twin-screw extrusion may be nutritional and functional alternatives to wheat-based products for consumers who cannot tolerate wheat gluten. Fig. 6 Microstructure of cross and longitudinal sections of commercial wheat pasta and an extrusion-cooked pasta-like product from pea flour. (a) and (b): cross and longitudinal sections, respectively, of commercial wheat pasta. (c) and (d): cross and longitudinal sections, respectively, of the pasta-like product prepared from pea flour by twin-screw extrusion (dough moisture, 32.0%; barrel temperature, 110 C; screw speed, 125 rpm; circular die opening, 1.09 mm dia). Fig. 7 Microstructure of cross and longitudinal sections of commercial wheat pasta and an extrusion-cooked pasta-like product from pea flour, surfaces treated with water. (a) and (b): cross and longitudinal sections, respectively, of commercial wheat pasta. (c) and (d): cross and longitudinal sections, respectively, of the pasta-like product prepared from pea flour by twin-screw extrusion (dough moisture, 32.0%; barrel temperature, 110 C; screw speed, 125 rpm; circular die opening, 1.09 mm dia). REFERENCES AACC American Association of Cereal Chemists. Approved Methods of the AACC, 9th ed. Methods 46-11A, 44-15A, and The Association: St. Paul, MN. Aguilera, J.M. and Kosikowski, F.V Soybean extruded product: A response surface analysis. J. Food Sci. 42: Aktan, B. and Khan, K Effect of high-temperature drying of pasta on quality parameters and on solubility, gel electrophoresis, and reversed-phase high performance liquid chromatography of protein components. Cereal Chem. 69: Anderson, R.A., Conway, H.F., Pfeifer, V.F., and Griffin, E.L.Jr Gelatinization of corn grits by roll and extrusion cooking. Cereal Sci. Today. 14: 4-8. AOAC Official Methods of Analysis, 15th ed. Association of Official Analytical Chemists, Washington, DC. Bailey, L.N., Hauck, B.W., Sevatson, E.S., and Singer, R.E Systems for manufacture of ready-to-eat breakfast cereals using twin-screw extrusion. Cereal Foods World. 36: Berset, C Color. Ch. 12 in Extrusion Cooking, C. Mercier, P. Linko and J.M. Harper, (Ed.), p American Association of Cereal Chemists, St. Paul, MN. Chiang, B.Y. and Johnson, J.A Measurement of total and gelatinized starch by glucoamylase and o- toluidine reagent. Cereal Chem. 54: Colonna, P. and Mercier, L Macromolecular modification of manioc starch components by extrusion cooking with and without lipids. Carbohydrate Polymers 3: Dexter, J.E., Matsuo, R.R., and MacGregor, A.W Relationships of instrumental assessment of spaghetti cooking quality to the type and the amount of material rinsed from cooked spaghetti. J. Cereal Sci. 3: Dexter, J.E., Kibom, R.H., Morhan, B.C., and Matsuo, R.R Grain research laboratory compression tester: Instrumental measurement of cooked pasta stickiness. Cereal Chem. 60: Diosady, L.L., Paton, D., Rosen, N., Rubin, L.J., and Athanssoulias, C Degradation of wheat starch in a single-screw extruder: mechano-kinetic breakdown of cooked starch. J. Food Sci. 50: Fellows, P Extrusion, Ch. 13 in Food Processing Technology: Principles and Practice, P. Fellows (Ed.), p Ellis Horwood Ltd., Chichester, UK. Kowlessar, O.D Dietary gluten sensitivity updated. J. Am. Diet. Assoc. 60(6): Likimani, T.A., Alvarezi, L., and Sosos, J.N The effect of feed moisture and shear strain on destruction Volume 64, No. 4, 1999 JOURNAL OF FOOD SCIENCE 677

8 Pasta-like Product From Pea Flour of Bacillus globigii spores during extrusion cooking. Food Micro. 7: Lomttillo, J.E., and Wolcott, J.M Process for producing pasta products. U.S patent 4,394,397. Lund, D.B Starch gelatinization. In Food Properties and Computer-aided Engineering of Food Process Systems. R.P. Singh and A.G. Medina (Ed.), p NATO ASI Series, Kluwer Academic Pub., London. Matsuo, R.R. and Irvine, G.N Note on an improved apparatus for testing pasta tenderness. Cereal Chem. 48: Matsuo, R.R., Malcolmson, L.J., Edwards, N.M., and Dexter, J.E A calorimetric method for estimating pasta cooking losses. Cereal Chem. 69: MINITAB Minitab User s Guide. Minitab Inc. State College, PA. Myers, R.H Response Surface Methodology. Allyan and Bacon, Boston. Rosenlund, M.L What is celiac disease? Clin. Pediatr. Phila. 9(12): 695. Smith, C., Van Megen W.V., Twaalfhoven, L., and Hitchcock, C The determination of trypsin inhibitor levels in foodstuffs. J. Food Agric. 31: Smith, O.B Why use extrusion? Paper presented at the 12th annual Central States Section of the American Association of Cereal Chemists, February 12-13, St. Louis, MO. Tyler, R.T., Youngs, C.G., and Sosulski, F.W Air classification of legumes. I. Separation efficiency, yield and composition of the starch and protein fractions. Cereal Chem. 58: Walsh, D.E Measuring pasta firmness. Cereal Sci. Today. 16: Wang, N., Bhirud, P.R., and Tyler, R.T Extrusion texturization of air-classified pea protein. J. Food Sci. 64(3): Wenger, M.L., and Huber, G.R Low temperature extrusion process for quick cooking pasta products. U.S patent 5, 059,439. Ms received 7/6/98; revised 1/29/99; accepted 2/10/99. We are grateful to the Food Processing Development Centre, Leduc, AB for use of the extruder. This project was funded by the Strategic Research Program of Saskatchewan Agriculture and Food, Regina, Saskatchewan, Canada. 678 JOURNAL OF FOOD SCIENCE Volume 64, No. 4, 1999