wt; .3? .5... .Q “fir noxiflfvl. .3- i: “may. 133., .lflistzvx . ,- . . . I. 1 a .3. V . A 115.....- . .Iifi , fin. 333.94. .. . V..;w!nv Lin .7 . THESE ’l 001’ ‘ . LIBRARY Michigan State Univergim 6/01 c:/CIRCIDateDue.p65—p. 1 5 THE EFFECT OF EXTRUDING WHEAT AT LOWER TEMPERATURES ON THIAMIN LOSS AND PHYSICAL ATTRIBUTES WHEN USING CARBON DIOXIDE GAS As A PUFFING AGENT By Abigail H. Schmid A THESIS Subnfifledto Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Food Science and Human Nutrition 2002 ABSTRACT THE EFFECT OF EXTRUDING WHEAT AT LOWER TEMPERATURES ON THIAMIN LOSS AND PHYSICAL ATTRIBUTES WHEN USING CARBON DIOXIDE GAS AS A PUFFING AGENT By Abigail H. Schmid Wheat flour with 0.3% (w/w) thiamin hydrochloride (vitamin B1) was extruded on a lab- scale extruder at lower temperatures and expanded using carbon dioxide (CO2) gas at 150 psi. Extrusion conditions based on preliminary work were barrel temperature profile of 40/40/50/70/80°C and screw speeds of 300, 350, and 400 rpm, at a flour feed rate of 3.6 kg/hr. These conditions were also repeated at 0 psi C02. High-temperature control samples were extruded at 40/60/90/130/150°C and screw speeds of 200, 250, and 300 rpm at 5.2 kg/hr. Dough moisture content was 22% in the control samples, and was 22 and 25% in the low-temperature samples. Maximum expansion ratios were 2.4 for low- temperature samples and 2.9 for high-temperature samples. Without C02, maximum expansion ratio was also 2.4. Expansion ratio increased with increasing screw speed, die product temperature, and energy input. Thiamin losses ranged from 10-16% in the control samples. With C02, thiamin losses were between 3-11% at 22% moisture, compared to 24-34% at 25% moisture. Without C02, thiamin losses were 0-1.5% at 22% moisture. Unlike typical high-temperature extrusion, thiamin loss in the low-temperature samples decreased with increasing screw speed. At 22% moisture using the extrusion set-up listed above, results indicate a potential for incorporating vitamins into raw flour prior to extruding at lower temperatures. DEDICATION To my wonderful family, Joe, Skip, and Sarah Schmid, for always being there and encouraging me to pursue my dreams. iii ACKNOWLEDGEMENTS First and foremost, I would like to thank my major professor, Dr. Kirk Dolan, for being an excellent mentor throughout this research project. It was a special honor being Dr. Dolan’s first graduate student. I would also like to thank the other members of my graduate committee, Dr. Wanda Chenoweth and Dr. Perry Ng, for their guidance and support. Special thanks to Dr. Maurice Bennink for the use of his fluorometer, Dr. Ng for the use of his extruder, and Richard Wolthuis for making the dies. Finally, I would like to thank Rujida Uthaisombut, Jeff Moore, and Mavis Tan for their technical support. Table 1.1. 2.1. 2.2. 2.3. 2.4 3.1.1. 3.1.2. 3.2.1. 3.2.2. 3.3.1. 3.3.2. 3.4.1. A.1. List of Tables Extent of Thiamin Destruction in Wheat Flour (13% Moisture, w/w) as a Function of Temperature and Mean Residence Times Screw Configurations for High-Temperature (Control) vs. C02- Injection Fluorometer Settings for the Varian SF-330 Spectrofluorometer Experimental Design for Data Collection Using Control and Carbon- Dioxide Screw Configurations Additional Experiments Using C02 Screw Configuration at Higher Temperatures The Effect of Die Product Temperature on Expansion Ratios for Samples in the C02 (at Lower Temperatures) and Control (at Higher Temperatures) Screw Configurations The Effect of Die Product Temperature on Expansion Ratios for Samples in the C02 ScrewtConfiguration at Higher Temperatures Average Bulk Density, WAI, and % WSI for CO2 (at Lower Temperatures) and Control (at Higher Temperatures) Screw Configurations Average Bulk Density, WAI, and % WSI for the C02 Screw Configuration at Higher Temperatures Mean Residence Times for CO2 (at Lower Temperatures) and Control (at Higher Temperatures) Screw Configurations Mean Residence Times for the C02 Screw Configuration at Higher Temperatures Average % Thiamin Losses and Mean Residence Times Without C02 at Lower Temperatures Physcial and Chemical Properties of Soft White Wheat Flour (Star of the West Milling Co.) Page 17 21 25 26 29 31 37 39 43 47 56 A.4. A.5.1. A52 A53. A.5.4. A.5.5. A.5.6. A.5.7. A.5.8. A.5.9. A.5.lO. A.5.ll. Reagent Amounts for the Construction of a Standard Curve (Fluorescence vs. Thiamin Concentration) Processing Data and Expansion Ratio, and Bulk Density Values for CO2 (Lower Temperature) and Control (Higher Temperature) Extrusion Conditions Processing Data and WAI, % WSI, SME and Mean Residence Time Values for CO2 (Lower Temperature) and Control (Higher Temperature) Extrusion Conditions Processing Data and Expansion Ratios, Bulk Densities, WAI, % WSI, SME, and Mean Residence Time Values for CO2 Extrusion at Higher Temperatures, 22% Moisture, Screw Speed of 300 rpm, and 0 psi C02 Processing Data and Average % Thiamin Losses for CO2 (Lower Temperature) and Control (Higher Temperature) Extrusion Conditions Processing Data and Average % Thiamin Losses for CO2 Extrusion at Higher Temperatures Averages and % CVs for % Thiamin in Feed Flour and Extruded Products on a Dry Weight Basis Averages and % CVs for % Thiamin in Feed Flour and Extruded Products on a Dry Weight Basis, Using C02 Extrusion at Higher Temperatures Example: Mean Residence Time Calculation, High Temperature (Control) at a Screw Speed of 200 rpm Results of MS Excel t-Test Comparing Mean Expansion Ratios at 0 and 150 psi CO2, for 3 Given Screw Speed and at 22% Moisture Results of MS Excel t-Test Comparing Mean % Water Solubility in the High- and Low-Temperature Samples at 22% Moisture, 0 psi C02, and 300 rpm Results of MS Excel t-Test Comparing Mean % Water Solubility in Low-Temperature Samples at 22% and 25% Moisture vi 65 7O 72 74 75 76 77 78 79 80 81 81 Figure 2.1. 3.1.1. 3.1.3. 3.1.4. 3.1.5. 3.3.1. 3.3.2 3.4.1 3.4.2 A.2.1. A.2.2. List of Figures Top View Of Screw Elements Aligned Inside the Barrel of an APV 1 9TC-25 Twin-Screw Extruder. Maximum Expansion Ratio for C02-Injected and High-Temperature (Control) Extrudates. Expansion Ratio Comparison for CO2 samples at 22% Moisture and 400 rpm Screw Speed With or Without C02-Injection. Average Expansion Ratio vs. Screw Speed Using the C02 (at Lower Temperatures) and Control (at Higher Temperatures) Screw Configurations. Average Expansion Ratio vs. Average Specific Mechanical Energy Using the C02 (at Lower Temperatures) and Control (at Higher Temperatures) Screw Configurations. Average Specific Mechanical Energy vs. Screw Speed Using the C02 (at Lower Temperatures) and Control (at Higher Temperatures) Screw Configurations. RTD Curves for Control (at Higher Temperatures) Screw Configuration at Screw Speeds of 200, 250, and 300 rpm at 22% Moisture. RTD Curves for CO2 (at Lower Temperatures) Screw Configuration at Screw Speeds of 300, 350, and 400 rpm at 22% Moisture, and 150 psi C02. Average % Thiamin Loss vs. Mean Residence Time for High Temperature Samples Using the Control and C02 Screw Configurations. Average % Thiamin Loss vs. Mean Residence Time for C02-Injected Samples (at Lower Temperatures) at 22 and 25% Moisture. The Effect of Salt Mass on Thiochrome Yield. The Effect of Potassium Ferricyanide Amount on Thiochrome Yield. vii Page 14 27 31 32 34 35 41 42 45 46 58 59 A.3. Standard Curves (Fluorescence vs. Thiamin Concentration) 61 Constructed in Duplicate on Two Different Days. viii Table of Contents List of Tables List of Figures Introduction 1. Literature Review 1.1. Extruded Food Products 1.1.1. High-Temperature Expanded Products 1.1.2. Carbon Dioxide-Expanded Products 1.2. Degradation in Extrusion 1.2.1. Introduction 1.2.2. Thiamin Analysis in Cereal Products 1.2.3. Effects Of Extrusion Conditions on Thiamin Loss 1.3. The Fortified Foods Trend 1.4. Research Objectives 2. Materials and Methods 2.1 . Raw Materials 2.2. Extrusion Conditions 2.1.1. Extrusion Set-up for High Temperature (Control) 2.1.2. Extrusion Set-up for C02-Injection 2.3. Expansion Ratio and Bulk Density 2.4. Water Absorption and Solubility Index 2.5. Specific Mechanical Energy Input 2.6. Mean Residence Time 2.7. Thiamin Analysis 2.8. Experimental Design 3. Results and Discussion 3.1. Expansion Ratio 3.2. Bulk Density, Water Absorption Index, and Water Solubility Index 3.3. Mean Residence Time 3.4. Thiamin Loss 4. Conclusions and Recommendations 4.1. Summary and Conclusions 4.2. Recommendations for Future Research Page vii HOOWQQ'JI-bww .— s—Ap—a 13 13 16 16 18 18 19 20 21 23 27 27 36 4O 49 49 52 Appendix 1 Appendix 2 Appendix 3 Appendix 4 Appendix 5 References 55 57 60 62 69 82 Introduction Extrusion is a continuous process primarily used for shaping and/or heating foods and polymers. In extrusion of foods, granular material (containing carbohydrates and proteins) is mixed with water, and transformed into a viscoelastic melt when subjected to heating, shearing, and conveying, via rotating screw elements located inside the extruder barrel (Rjzvi et a1., 1995). A forming die is located at the end of the barrel where the doughy mass is pushed through and shaped. Extruded food products include pasta, confectionary products, texturized vegetable protein, pet foods, puffed snack foods and ready-to-eat cereals. Some of the advantages of extrusion, that also give the process longevity in the food industry, are versatility, high productivity, high product quality, and energy efficiency (Harper, 1981). One unique aspect of extrusion is the ability to produce puffed products. Often the forming die is conical-shaped and the narrowest point is at the die exit. As the flow of dough is restricted at the die,'hightpressures are generated (600-900 psi) and at high barrel temperatures (140-1 80°C) the water inside the dough becomes superheated. When exiting the die into atmospheric pressure, the water converts to steam and expands and cools the extruded product as it moves through the dough and escapes to the atmosphere. Mercier and Feillet (1975) reported that the most influential variables on expansion ratio (diameter of extrudate/diameter of die) are extrusion temperature and feed material moisture content. They found that in a twin-screw extruder, maximum expansion ratio was obtained for cornstarch at 14% moisture content and temperatures between 180- 200°C in the tested moisture content range of 10.5 to 28.5% (w/w). One disadvantage of extruding at high temperatures and low moisture contents is the loss of vitamins due to the breaking of chemical bonds during mixing, or the loss of vitamin stability due to heat (ijrck and Asp, 1983). For thiamin loss during extrusion, Beetner-et al. (1974) reported 46% average loss and an increase in vitamin loss as temperature and screw speed increased. In extruded wheat flour, Guzman-Tello and Cheftel (1987) found a decrease in thiamin loss as the moisture content increased. Due to the significant loss of thiamin during extrusion, ready-to-eat cereals are fortified with thiamin and other vitamins post- extrusion, such as by spraying onto the surface of the extrudate (Burns et al., 2000). There have been some investigations of lower-temperature (near 100°C) extrusion of cereals by injecting supercritical carbon dioxide (Mulvaney and Rizvi, 1993; Lee et al., 1999) or carbon dioxide gas (Ferdinand et al., 1990, 1992) as expanding agents. Vitamin degradation was not investigated. Other researchers have hypothesized that if the harsh conditions of the extrusion process are reduced, then vitamin loss will decrease (Lee et al., 1999; Sokhey et al., 1996). Killeit (1994) reported that vitamin degradation in extruded foods increased with increasing temperature, screw speed, and specific energy input, and with decreasing throughput, moisture, and die diameter. Carbon dioxide injection is an inexpensive way of incorporating nucleated gas bubbles to expand the product at lower extrusion temperatures (Mulvaney and Rizvi, 1993). If vitamin loss is reduced at lower temperatures, then the post-extrusion addition of vitamins may become unnecessary. Thus, C02-injection coupled with the addition of vitamins to feed material, has the potential of lowering processing costs, which would benefit breakfast cereal and snack food industries. 1. LITERATURE REVIEW 1.1. Extruded Food Products The food industry began using extrusion approximately seventy years ago after borrowing the application from the plastics industry, which used extruders to shape products by forcing the plastic melt through an opening. Today, screw-type extruders are used to manufacture food and feed products. The screw-type design uses flighted screw elements on a rotating shaft to convey the material along the extruder barrel toward the exit die. Screw extruders can be categorized as single-screw, co-rotating twin-screw, or counter-rotating twin—screw. The co-rotating twin-screw extruder is the most efficient because it has the greatest capacity for mixing products unifonnly (Harper, 1988). In addition to acting as a former, an extruder can also be used to cook the shaped products. The cooking action takes place inside the extruder. barrel through the combination of dissipated mechanical energy from the rotating screws plus optional added thermal energy by injected steam, or electrically heated rods or jackets. Maximum temperatures toward the exit die typically reach between 140°C and 190°C, but due to the movement of the material by the screws, the residence times at these high temperatures is much shorter than the total residence time inside the entire barrel. Because of this, extrusion cooking can be classified as a high-temperature/short-time (HTST) process. Similar to the HTST heat treatment of raw milk, this process is effective in the destruction of spoilage enzymes and microorganisms (Harper, 1981). In addition to improving product quality, some of the other advantages of using extrusion in food processing include the ability to produce a variety of products, continuous processing, and energy efficiency (Hauck and Huber, 1989). These advantages are becoming even more important as the consumer demand for new food products increases. 1.1.1. High-Temperature Expanded Products In the extrusion cooking process, molecular changes, such as the disruption of starch granules and protein denaturation, are likely to occur. These changes typically involve the conversion of a starchy and/or proteinaceous raw material into a viscoelastic melt. This transformation is important to the expansion of products since the viscoelastic properties of the melt allow for air cells to be retained, resulting in a rigid, porous, structure upon exiting the die (Kokini et al., 1992). Starch gelatinization (often measured by loss of birefringence) results as the starch molecules are heated and mixed with water inside the extruder barrel. For wheat flour, Chiang and Johnson (1977) reported that increased extrusion temperatures increased starch gelatinization when the tested moisture content range was between 18 and 27% (w/w). For extruded cornstarch, Bhattacharya and Hanna (1987) found that expansion ratio increased and starch gelatinization was greater when barrel temperatures increased fi'om 116°C to 166°C. According to their study, expansion ratio increased with decreasing moisture contents. They attributed this to higher moistures possibly having a decreased effect on product temperatures, thus causing expansion ratio to decrease. In summary, conventional puffing is typically done at high temperatures and low moisture contents to create ideal conditions for water vapor flash-off and starch gelatinization. Expansion ratio, bulk density, water absorption index (W AI), and water solubility index (WSI) are physical characteristics of expanded products often used to represent extrudate quality. Expansion ratio measures how much the dough expands after exiting the die. Bulk density determines the weight of product per unit volume and is related to expansion ratio, since the greater the expansion ratio the lower weight per volume the product will occupy. A low bulk density is desirable when trying to package expanded products in a cost effective manner. In the ready-to-eat breakfast cereal industry, WAI and WSI are directly related to “bowl life”, or the amount of time it will take for the product to become soggy in milk. The lower the WAI and WSI the longer the “bowl life” which will most likely result in a crispy and crunchy product throughout the breakfast eating experience (Sokhey et al., 1996). 1.1.2. Carbon Dioxide-Expanded Products To date, limited research has been conducted on the use of injected carbon dioxide to expand extruded products. The idea of using carbon dioxide injection in food extrusion was also borrowed from the plastics industry, as gases were traditionally used to produce plastic foams (Ferdinand et. al., .1990). The two types of carbon dioxide injection used by past researchers in the extrusion of cereal products have been supercritical fluid carbon dioxide and pressurized carbon dioxide gas (Lee et al., 1999; Sokhey etal., 1996; Rizvi et al., 1995; Mulvaney and Rizvi, 1993; Ferdinand .et al., 1990, 1992). In both cases, at lower barrel temperatures, carbon dioxide injection of a fluid or gas was used to inject nucleated bubbles in the viscoelastic melt (Mulvaney and Rizvi, 1993). In the case of the supercritical fluid, the fluid changes to a gas as the pressure reduces to atmospheric conditions on exit fiom the die. In the supercritical C02 studies, process optimization was investigated by changing screw configuration, pressure profiles, barrel temperatures, screw speeds, moisture contents, and by adding texturizing agents (whey protein isolate and non-fat dry milk). To look at whether changing processing variables had significant effects on physical attributes and chemical properties, extrudates were characterized for starch gelatinization, expansion ratio, breaking stress (amount of force required to break an extrudate), water absorption index, and water solubility index. In the C02 gas-injection studies, similar properties of extrudates were characterized, but for different extrusion conditions in which greater emphasis was placed on minimizing barrel temperatures (<100°C) in the cooking zone. The effect of extrusion variables on specific mechanical energy (SME) input was also investigated for supercritical and C02 gas injection (Mulvaney and Rizvi, 1993; Lee et al. 1999; Ferdinand et a1. 1990, 1992). SME has been defined as the amount of mechanical energy required to process 1 kg of flour and water input, and is an important parameter in estimating manufacturing costs (Bhattacharya and Choudry, 1994). In high-temperature expanded products, puffing is maximized under harsh extrusion conditions such as high temperature and low moisture contents (Mercier and Feillet, 1975). Because of this, some of the benefits that carbon dioxide-injected extrusion offers include reduced wear on the barrel, lower energy costs, and the ability to potentially add heat-sensitive ingredients like flavors, colors, and vitamins in with the raw material (Rizvi et al., 1995). In addition, carbon dioxide is a relatively inexpensive expanding agent. Oxygen is also a low-cost expanding agent, but it has not been investigated in food extrusion, due to its flammability and oxidizing capabilities. 1.2. Thiamin Degradation in Extrusion 1.2.1. Introduction Thiamin, also called Vitamin B1, is an essential, water-soluble vitamin that is important in carbohydrate metabolism, energy production, proper function of the nervous system, and in the prevention of beriberi, an endemic disease in parts of Asia where polished white rice is a major dietary staple (Gubler, 1991). Beriberi is categorized into two main types, wet beriberi and dry beriberi. Wet beriberi symptoms include the accumulation of fluids in the feet, legs, and ankles, and is also referred to as edema. This edema may also cause the heart to enlarge and lead to congestive heart failure. Dry beriberi symptoms include muscle emaciation and nerve abnormalities. Other general symptoms include anorexia and difficulties in walking (Gubler, 1991). The chemical structure of free thiamin is pyrimidine plus thiazole, attached by a methylene bridge. The coenzyme form is thiamin pyrophosphate, and is formed in the body when a pyrophosphate group is added to the structure of free thiamin (Rindi, 1996). The Daily Value (nutritional recommendation based on Reference Daily Intakes or the former US. Recommended Dietary Allowances) for thiamin is 1.5 mg/day (Food and Drug Administration, 1994). Dietary sources .of thiamin include fortified ready-to-eat whole-grain cereals, enriched bread and flour, and pork, legumes, and dairy products (Rindi, 1996). Stability characteristics for thiamin have been reported as being readily destroyed at temperatures near 100°C and in solutions with a pH above 6.2 (Mulley et al., 1975a). Thiamin instabilities are due to the cleavage of the methylene bridge into pyrimidine and thiazole fragments when heated in alkaline solutions (Dwivedi and Arnold, 1973). 1.2.2. Thiamin Analysis in Cereal Products The categories of methods available for the analysis of thiamin include chemical, microbiogical, animal, and physical. For food and feed products, thiamin content can be measured rapidly and economically using a chemical method (Association of Vitamin Chemists, 1966). The thiochrome method is the standard chemical method used by the Association of Official Analytical Chemists (AOAC, 1995) and the American Association of Cereal Chemists (AACC, 2000). When thiamin is oxidized, it forms thiochrome, which fluoresces under UV light. A fluorometer is then used to measure the amount of fluorescence, which is linearly proportional to the amount of thiamin under standard conditions and when other fluorescing compounds are not present (Mulley et al., 1975b). When thiamin loss was investigated in extruded wheat flour, Guzrnan—Tello and Chefiel (1987) used the thiochrome method for determining thiamin concentration. Pharn and Del Rosario (1986) and Maga and Sizer (1978) also determined thiamin content in extruded legumes and extruded potato flakes, respectively, with the thiochrome method. 110 and Berghofer (1998) and ' Beetner et a1. (1974) used microbiological methods for measuring thiamin in extruded corn grits. Although not listed as an official method by AOAC or AACC for thiamin analysis, Toma and Tabekhia (1979) and Kamman et a1. (1980) have reported that high performance liquid chromatography (HPLC) is also a rapid and accurate method for measuring thiamin concentration in foods. In these two studies, HPLC and the thiochrome method were used to measure thiamin content in the same products. When thiamin concentrations of the two methods were compared, they found no difference in values. Since either method was acceptable and the thiochrome method has been reported to be accurate within 5%, the thiochrome method was used in the present study to measure thiamin in raw flour and extruded products (Labuza and Riboh, 1982). 1.2.3. Effects of Extrusion Conditions on Thiamin Loss Thiaan is naturally present in cereal grains, but thermally and mechanically degraded during processing, such as in milling and extrusion. Guzman-Tello and Chefiel (1987) extruded soft wheat flour mixed with thiamin at identical conditions for four different initial concentrations of thiamin. They plotted log(initial B1 concentration) vs. log(final B1 concentration). The slope was 1.03, indicating that thiamin degradation in extruded wheat flour follows virtually a first-order kinetic reaction. As shown in Table 1, they found the destruction rate constant (k) of thiamin to be a function of die product temperature and mean residence time (mean time the raw material spends inside extruder barrel): Table 1.1. Extent of Thiamin Loss in Wheat Flour (13% Moisture, w/w) as a Function of Temperature and Mean Residence Times' Treatment # Product Screw Mean Residence % Thiamin k x 107 Temperature Speed Time in Heating Lossb (per Just Before (rpm) Zone of Extruder (dry basis) sec.) Die (C)° (seconds) 1 131 100 42 11.5 (i0.l) 3.23 2 145 100 41 21.5 ($0.3) 6.86 3 160 100 42 30.0 (3:09) 9.34 4 176 100 40 42.5 (ilJ) 14.4 5 159 100 43 27.0 ($1.3) 7.92 6 159 125 40 41.1 (3:3.5) 13.7 7 160 150 34 47.2 ($4.7) 19.5 “ Adapted from Guzman-Tello and Chefiel, 1987 b Experiments conducted in duplicate. Standard deviations are given in parentheses. In comparing treatments #3 to #7, it appears that at constant die temperatures, increasing screw speed from 100 to 150 rpm increased the percent thiamin loss by approximately 17%. With increasing screw speed, the amount of time the product spent inside the barrel decreased (34 vs. 42 seconds), indicating that shear effects may have contributed more to thiamin loss than thermal effects. 110 and Berghofer (1998) reported thiamin losses in the extrusion of corn grits from 67-100%, when barrel temperatures ranged from l40-200°C, screw speeds from 65-81 rpm, and feed moistures from 11.8-14.2% . For non-enriched white flour, thiamin losses were 58% when the temperature at the die was 197°C, screw speed was 200 rpm, and feed moisture was 14.6% (Hankansson et al., 1987). Similar studies on thiamin degradation in extrusion indicated that thiamin .loss decreased as energy input decreased or feed material moisture content increased (Asp and Bjdrck, 1989; Pharn and Del Rosario, 1986; 110 and Berghofer, 1998; Killeit, 1994). These researchers suggested that an increase in moisture content reduces the dissipation of mechanical energy, resulting in a decrease in thiamin loss. 1.3. The Fortified Foods Trend Consumer demand for new, convenient products continues to grow rapidly (Sloan 2001). Based on consumer trends, 3 type of product worth exploring would be one that offers health benefits at a reasonable cost. Sloan (2000) reported that “nine out of ten shoppers now believe that healthy eating plays a role in disease prevention”. Sloan also reported that fortified foods are a $50 billion dollar industry and two-thirds of those that grocery shop in the US. are buying more fortified foods. Ready-to-eat cereals, a type of extruded product, have traditionally been fortified, but with additional equipment to spray vitamins onto the product post-extrusion. The additives (vitamins, colors, flavors, and other nutrient additives) are added post-extrusion due to the harsh conditions inside the extruder barrel that degrade the chemical structure of some additives. Because of this, a 10 challenge remains to the extruded foods industry to modify the extrusion process where more nutrients are retained throughout processing, thus eliminating the need for an additional fortification step. One approach to addressing this challenge would be to add the nutrients to the raw material, thereby resulting in a more uniform distribution of the nutrients than that resulting fiom the spraying method. 1.4 . Research Objectives Extruding foods at lower temperatures has potential for lowering processing energy costs and decreasing the loss of heat- and/or shear-labile nutrients. However, lower temperatures would decrease expansion normally caused by steam flashing off the extrudate, and would result in a less appealing product. Injection of C02 gas may decouple the dependence of expansion on high temperatures, and allow expansion at lower temperatures. A few studies investigated the potential of CO2 gas extrusion but these studies did not report vitamin loss. Therefore, the research objectives are as follows: Objectives 1. To find a screw configuration, moisture content range, screw speed range, die geometry, CO2 pressure and barrel temperature profile near 100°C that will sufficiently melt and puff the extruded wheat dough. 2. To investigate the effects of moisture content, screw speed, and C02 pressure on the extruder specific mechanical energy input and mean residence time, and on the physical characteristics of expansion ratio, water absorption index, water solubility index, and bulk density. 11 3. To compare thiamin loss in products puffed at lower temperatures (near 100°C) to thiamin loss in products puffed at conventional temperatures (160°C). 12 Materials and Methods “Images in this thesis are presented in color.” 2.1. Raw Materials Michigan soft white wheat pastry flour (Star of the West Milling Co., Frakenmuth, MI) with a moisture content of approximately 13% was blended with 0.3% (w/w) food-grade thiamin hydrochloride (Spectrum laboratory Products, Inc., Gardena, CA). Using a twin-shell dry blender (Patterson-Kelley, East Stroudsburg, PA), the flour and thiamin were mixed for 40 minutes, the minimum time to ensure adequate distribution of thiamin (coefficient of variance < 5%). The blender consisted of an inner mixing bar that held 4 cm long splines. The tip speed of one spline was measured with a manual tachometer as 5.9 m/s. According to Patterson-Kelley, high and medium speed blenders usually have tip speeds of 17 and 8.5 rn/s, respectively. The, mixing procedure and the chemical and physical properties of flour are outlined in Appendix 1, Table All 2.2. Extrusion Conditions An APV (Grand Rapids, MI) MPl9TC-25 co-rotating and intermeshing twin- screw extruder was used to extrude eachsample in duplicate. The diameter of each barrel is 19 mm, thus making the length-to-diameter ratio of the extruder 25:1. The size of the extruder used in the present study was much smaller than an industrial—size extruder. Along the length of the barrel and die, there are five zones in which the temperature is controlled via thermocouples, electrical heating elements, and water- cooling jackets. The direct current motor supply provides power to turn the shafts and screws up to a maximum screw speed of 500 rpm. A control display panel allows the 13 operator to monitor process variables such as the torque required to turn the screws (expressed as a percentage of the maximum torque), screw speed (rpm), temperature of the product and barrel (°C), and die pressure (psi). Die pressure was measured using a pressure transducer (Dynisco, Model # EPR3-3M-6) located 7 mm before the die entrance. The product temperature inside the die was measured by hand-inserting a T- type needle thermocouple (Cole-Parmer, Vemon-Hills, IL) into the die hole during extrusion. The exit dies both have circular openings of 3 mm and lengths of either 6 mm or 12 mm. “a.“ -R. punun‘w I, . 9 .luiILL adScfus § In I ghmglc-l .czltl 591w» 5 t ‘1"“1‘ _ —..._— 4 __ _ 1 .u lfiixing Puddles 9.. Figure 2.1. Top View of Screw Elements Aligned Inside the Barrel of an APV MPl9TC-25 Twin-Screw Extruder. Figure 2.1 shows how the screw elements appear inside the barrel. The helical metal rib around the screw is called a flight, and lead typically refers to the axial distance 14 between flights (Harper, 1981). In this case, a twin-lead screw has the geometry to fit two helices, whereas the single-lead fits one helix. The twin lead geometry is most ideal for conveying and the single lead geometry is better for compressing and heating the dough. When the screw elements are next to one another in Figure 2.1, twin-lead screws are perpendicular to each other, and the single-lead screws are fully intermeshed. They are aligned in this manner to ensure optimal mixing along the barrel. The mixing paddles have a greater capacity for shearing and move along the barrel in either a clockwise (forward) or counterclockwise (reverse) motion. The paddle orientation can be 30, 60, or 90°. According to APV specifications, 30° has the greatest conveying efficiency because it operates at a low degree of fill, and 90° has a lower conveying efficiency by operating at a higher degree of fill. The rate of feed material and deionized water into the extruder are controlled respectively by a K-Tron K2M twin-screw volumetric feeder (K-Tron Corp., Pitrnan, NJ) and E2 Metripump positive displacement metering pump (Bran & Luebbe, Northampton, UK). The twin-screw volumetric feeder is equipped with a safety feeder, variable speed drive, and digital control. Prior to sample collection, the extruder was run for at least five minutes at steady state (constant torque, die pressure, and die temperature). After recording the process variables (screw speed, % torque, temperature profiles, die pressure, screw configuration, feed and water injection rates, and product temperature at the die), samples were collected for five minutes and held at room temperature in an area of subdued light prior to drying. Drying took place in a convection fan oven at 75°C for 16 hours, to a moisture content of 3.5 to 7%. After drying, the extrudates were stored in a light-resistant, sealed, 15 polyethylene bag at 1°C until further analysis. About 30 g of extruded material were ground using a Udy Cyclone Mill (Udy Corp., Fort Collins, CO) with a 0.5 mm screen. Moisture content was determined in duplicate on a 2.5-g ground sample, by heating at 130°C for 10 minutes using a Sartorius MA-30 moisture analyzer (Gtiettingen, Germany). 2.2.1. Extrusion Set-up for High Temperature (Control) The screw configuration listed in Table 2.1 was one that had been used in preliminary experiments to expand extruded cereal products at high temperatures. Extrusion parameters included a constant temperature profile of 40/60/90/ 130/ 150°C (increasing temperature toward die), die geometry of 3 mm diameter x 12 mm length, feed rate of 5.2 kg/hr, and water injection rate of 0.69 kg/hr. The moisture content inside the barrel was 22% for the above feed rate and water injection rate. Three constant screw speeds of 200, 250, and 300 rpm were used. The above parameters had been tested in preliminary runs and identified as conditions that produced optimal puffing (expansion ratios > 2.0). 2.2.2. Extrusion Set-Up for CO2-lnjection. The screw configuration for CO2-injection listed in Table 2.1 was based on preliminary experiments for producing a puffed product with CO2 gas as the expanding agent. Placing mixing paddles at the point of C02-injection in order to prevent the backflow of CO2 gas firrther optimized this screw configuration. Ferdinand et a1. (1990) used a similar mixing paddle set-up when the point of C02 injection was 165 m away from the die entrance. In the present study, the point of CO2-inj ection was 110 mm away from the die entrance. The point of C02-injection had been tested in previous studies 16 using the same APV extruder to produce expanded products. The optimal extrusion parameters for puffing were determined in a manner similar to the control samples. Each parameter (screw speed, moisture content, C02 pressure, and die geometry) was changed one at a time until a uniformly puffed product was obtained without barrel temperatures exceeding 100°C. These optimal parameters included a constant set temperature profile of 40/40/50/70/80°C, die geometry of 3 mm diameter x 6 mm length, C02 pressure of 150 psi, and feed rate of 3.6 kg/hr. The water injection rates were 0.36 kg/hr and 0.52 kg/hr resulting in feed material moistures of 22% and 25%, respectively. The screw speeds were 300, 350, and 400 rpm. Table 2.1. Screw Configurations for High-Temperature (Control) vs. CO2-Injection High Temperature CO2-Injection 8Da Twin Lead 7D Twin Lead 7x30° Forward Mixing Paddle 7x30° Forward Mixing Paddle 8D Twin Lead 5D Twin Lead 3x60° Forward Mixing Paddle 1D Single Lead 3x30° Reverse Mixing Paddle 3x60° Forward Mixing Paddle 2D Single Lead 3x60° Reverse Mixing Paddle 4x60° Forward Mixing Paddle 2D Single Lead 3x30° Reverse Mixing Paddle 3x60° Forward Mixing Paddle 2D Single Lead 4x60° Reverse Mixing Paddle 4D Twin Lead 1D Single Lead “D = 19 mm 17 2.3. Expansion Ratio and Bulk Density After drying for 16 hours, calipers were used to measure the diameter of extruded products. Per sample, five strands were randomly selected and three measurements were taken along the product. To calculate expansion ratio, the average of 15 measurements was divided by the die diameter of 3 mm. Bulk density was determined by filling a 200 mL graduated cylinder with extruded pieces that were approximately 0.5 cm long. The graduated cylinder was tapped gently on a flat surface before weighing. The following equation was used to determine bulk density: Bulk Density = g of sample per 200 mL 200 mL For each sample, bulk density was measured in triplicate. 2.4. Water Absorption Index (WAI) and Water Solubility Index (WSI) The determinations of WAI and WSI were based on the method used by Anderson et a1. (1969) where 2.0 g of ground extrudate is suspended in 20 mL of water at 30°C, shaken for 30 minutes and centrifuged at 3000 rpm (1075 x g) for 15 minutes. The supernatant was dried in a convection oven at 130°C for 2 hours. Taking two measurements per sample, WAI was calculated as the weight of sediment (or gel) Obtained per gram of dry sample or: WAI= (weight of sediment + centrifige tube) - (weight of centrifuge tube) sample dry weight WSI was calculated as a percentage of soluble material using the following equation: % WSI= weight of dish + dried supern3a_t_ant) - (weight of dish) x 100 sample dry weight . l8 2.5 . Specific Mechanical Energy Input Specific mechanical energy (SME) is a measure of the viscous energy dissipation per unit mass of dough (Mason and Hoseney, 1986). This energy originates from the direct current motor supply which provides the required torque to turn screws at a given screw speed. A portion of total power supplied to the shaft is used to push dough along the barrel against back-pressure generated at the die (Mohamed et al., 1990). The remaining power is transferred into the dough via fluid fiiction. Therefore, specific mechanical energy (SME) is calculated as the total power supplied to the shaft (PW), minus power to convey the dough (APQ), divided by the mass flow rate: SME (kJ/kg)=i (1%) “Wk—g?) (Pw _ APQ) where Ev = —— 1000 Pw = (0.044)(%torque)(N), J/s U as given by the manufacturer N = screw speed, rpm AP = die pressure, Pa . m Q = volumetrrc flow rate, — , m3/s m = mass flow rate of dough, Infeed “on, + r'nm3r , kg/s p = dough density, kym3 For the dough density measurement, a sample of dough was taken near the entrance of the die. The dough sample was wrapped tightly in plastic wrap and weighed. After recording the weight, the dough was placed into a 100 mL graduated cylinder filled with 30 mL deionized water. Dough volume was recorded as the volume of water displaced l9 by the dough. Dough density was calculated as the weight of dough (g) divided by the volume of dough (mL) and then units were mathematically converted from g/mL to kg/m3. Average SME was reported as the average between two replicate extrusion conditions. 2.6 . Mean Residence Time Due to axial mixing within the extruder barrel, there is a distribution of residence times for dough particles. Mean residence time is affected most strongly by screw configuration, screw speed, and feed rate (Altomare and Ghossi, 1986). Solid Red 40 dye in the amount of 0.3g acted as a traceable color indicator and was poured by hand instantaneously where feed flour enters the extruder. Time at which red color first appeared in the extrudate exiting the die was measured. Extrudate strands were cut every five seconds until no obvious color appeared at the exit die. A strand was collected prior to color injection to represent the color value at time zero. A coffee grinder (Sunbeam Corporation, Maitland, FL) was used to reduce the particle size of the strands. Each sample in the amount of 1.6 g was then placed into a round disc with a 3.5-cm diameter and 0.5-cm height. A HunterLab D25 L color meter (Hunter Associates Laboratory, Reston, VA) measured the a*-va1ue (redness) of the ground samples. Standard number C2-30954 was the white plate used to standardize the color meter. Redness values were recorded and were used to represent the intensity (C(t)) of red color at exit time t. Mean residence time was calculated as a weighted average (Levenspiel, 1999): = Zr C(t)At ZC(t)At Mean residence time (seconds) 20 Microsoft Excel was used to construct a residence time distribution curve of E(t) C (t) ormal'zed concentrat'on vs.t seconds , where E t = . (N 1 l ) ( ) () 2 C(t)At 2.7 . Thiamin Analysis Thiamin concentration determination was based on the 1995 AOAC “Official Method 953.17 Thiamin (Vitamin Bi) in Grain Products, Fluorometric (Rapid)”. In this method, potassium ferricyanide oxidizes thiamin to thiochrome, which fluoresces under UV light. The spectrofluorometer type was SF-330 from Varian (Palo Alto, CA) and methylacrylate sample cuvettes were used. The fluorometer was standardized with quinine sulfate at an excitation wavelength of 343 nm and emission wavelength of 459 nm. The wavelength range for measuring thiamin concentration was excitation wavelength of 373 nm and emission wavelength of 410 nm. The fluorometer settings used for analyzing all samples are listed in Table 2.2: Table 2.2. Fluorometer Settings for the Varian SF-330 Spectrofluorometer Light Source Xenon Lamp Excitation slit 5 nm Emission slit 10 nm Sensitivity x 1/ 10, samples x 10, blanks Selector x 1 Modifications to the AOAC method were tested and implemented to improve the accuracy, precision, and efficiency of thiamin analysis. According to the AOAC method, sodium chloride should be added in the final oxidation step to act as a drying agent. However, Brubacher et al. (1985) reported that thiochrome yield may be reduced by 15- 21 19% if sodium chloride is present during oxidation and should therefore be added after oxidation is complete. Brubacher et al. (1985) also reported that when in excess, potassium ferricyanide can destroy thiamin. Based on this research, we did our own study and found results similar to Brubacher et al.’s (1985). The addition of NaCl and potassium ferricyanide was then adjusted in the oxidation step in order to optimize thiochrome yield (Appendix 2, Figures A21 and A.2.2). Using the exact quantities and volumes of reagents listed in the AOAC method, thiamin concentration is calculated using the following equation: (I-b) (s-d) pg thiamin in 5 ml assay solution = In this equation, I is the fluorescence reading for the oxidized sample, b is the reading for the non-oxidized sample, S is the reading for the oxidized standard, and d is the reading for the non-oxidized standard. This method uses a standard concentration that covers the entire range of sample solutions and assumes that there is a linear relationship between thiamin concentration and fluorescence. In the present study, a standard curve of fluorescence vs. thiamin concentration was constructed to confirm a linear relationship between the two (Appendix 3, Figure A.3.1). Since the day-to-day standard curve variability was very low when slopes were compared (<2% variability), a standard curve was not run on each day of analysis. Instead, one was constructed periodically, and averages of the y-intercept and slope values were used in the final calculation for percent thiamin. Percent thiamin was calculated in triplicate for each sample on a dry weight basis using the following equation: 22 (I-b) - z x (F.)(F.)(10*) m % thiamin = x 100% I and b are the same as above and z and m are the average y-intercept and s10pe, respectively, from the standard curves. F 1 is the dilution factor that occurs in the extraction step and F2 is the dilution factor that occurs in the oxidation step. Wrepresents the dry weight of the sample (g). The equation for calculating percent thiamin loss was: % thiamin loss =[ —£] x100 0 Where P0 is the average percent thiamin in feed flour (dry basis, based on four measurements) and P; is the average percent thiamin in the extruded product (dry basis, based on three measurements). This modified method for thiamin analysis is presented in detail in Appendix 4. 2.8 Experimental Design In the present study, the independent process variables included moisture content (%), die geometry (mm), screw speed (rpm), temperature profile (°C), carbon dioxide pressure (psi), screw configuration, feed rate (kg/hr), and initial thiamin concentration (0.3% w/w). The dependent variables were die pressure (psi), temperature of product at die (°C), specific mechanical energy (kJ/kg), expansion ratio, bulk density (g/mL), water absorption index, water solubility index, mean residence time (seconds), and % thiamin loss. Table 2.3 shows the experimental design for the control and CO2 screw configurations. In addition to screw configuration, temperature profile, and moisture content, the constant independent variables for high temperature controls were feed rate (5.2 kg/hr), initial thiamin concentration (0.3% w/w), and die geometry (3 mm diameter x 12 mm length). A range of screw speeds was used to see if there was a significant 23 difference between dependent variables and changes in screw speed. In samples 7-30, screw configuration, temperature profile, feed rate (3.6 kg/hr), initial flour thiamin concentration (0.3%), and die geometry (3 mm diameter x 6 mm length) were the constant independent variables. A range of moisture contents, screw speeds, and C02 pressures were used to see if there was a significant difference between the dependent variables and different extrusion conditions. Samples 7-12 and 19-24 were also considered controls to show the effect of C02 pressure on the dependent variables when no CO2 was injected. For each extrusion condition, replicate runs were conducted on different days. For example, samples 1, 3, and 5 were run on one day, and 2, 4, and 6 were run on a different day. Listed in Table 2.4 is a set of additional experiments that were run after samples 1-30 were collected. - The purpose ‘of adding these two experiments was to investigate whether or not the C02 extrusion set-up (CO2 screw configuration, 3 mm x 6 mm die, and 3.6 kg/hr feed rate) could produce quality expanded-products using the higher (control) temperature profile, 0 psi. CO2, screw speed of 300 rpm, and 22% moisture content. Another reason was to see how these extrusion conditions affected mean residence time and thiamin loss, and also to compare these mean residence time and thiamin loss values to those measured in samples 5 and 6 (Table 2.4 compared to Table 2.3). 24 Table 2.3. Experimental Design for Data Collection Using Control and Carbon- Dioxide Screw Configurations 'b Sample Screw Moisture Temperature Profile 902 Screw Configu_ration Content (3) Pressure Speed (psi) (rpm) 1 Control 22% 40/60/90/130/150 N/A 200 2 Control 22% 40/60/90/130/ 150 N/A 200 3 Control 22% 40/60/90/130/150 N/A 250 4 Control 22% 40/60/90/130/150 N/A 250 5 Control 22% 40/60/90/130/150 N/A 300 6 Control 22% 40/60/90/130/150 N/A 300 7 CO2 22% 40/40/ 50/ 70/80 0 300 8 CO2 22% 40/40/50/70/80 0 300 9 CO2 22% 40/40/50/70/80 0 350 10 C02 22% 40/40/50/70/80 0 350 1 1 CO2 22% 40/40/50/70/80 0 400 12 CO2 22% 40/40/50/70/80 O 400 13 C02 22% 40/40/50/70/80 150 300 14 . CO2 22% 40/40/50/70/80 150 300 15 CO2 22% 40/40/50/70/80 150 350 16 CO2 22% 40/40/50/70/80 150 350 17 CO2 22% 40/40/50/70/80 150 400 18 C02 22% 40/40/ 50/70/80 150 400 19 C02 25% 40/40/50/70/80 O 300 20 CO2 25% 40/40/50/70/80 O 300 21 CO2 25% 40/40/50/70/80 O 350 22 C02 25% 40/40/50/70/80 O 350 23 C02 25% 40/40/50/70/80 0 400 24 CO2 25% 40/40/50/70/80 O 400 25 CO2 25% 40/40/50/70/80 150 300 26 C02 25% 40/40/50/70/80 150 300 27 CO2 25% 40/40/50/70/80 150 350 28 C02 25% 40/40/50/70/80 150 350 29 CO2 25% 40/40/50/70/80 150 400 30 CO2 25% 40/40/50/70/80 150 400 -3 Die geometry = 3 mm diameter x 12 mm length, samples 1-6 = 3 mm diameter x 6 mm length, samples 7-30 b Feed rate = 5.2 kg/hr, samples 1-6 = 3.6 kg/hr, samples 7-30 25 Table 2.4. Additional Experiments Using CO2 Screw Configuration at Higher Temperatures 'b Screw Moisture Temperature Profile C_02 Screw Sample ConfiguLation Content (‘3) Pressure Speed (Pail (rpm) 31 CO2 22% 40/60/90/130/150 0 300 32 CO2 22% 40/60/90/130/150 0 300 ‘ Die geometry = 3 mm diameter x 6 mm length b Feed rate = 3.6 kg/hr 26 Results and Discussion 3.1. Expansion Ratio r. »- _ ~. .. .~ W. ~ .. 4. .3‘3 . Figure 3.1.]. Maximum Expansion Ratio for C02-Injected and High- Temperature (Control) Extrudates. Figure 3.1.1 shows the maximum expansion ratio for the CO2-injected and control samples. For the CO2-injected extrudates, maximum expansion ratio was obtained at a screw speed of 400 rpm, 22% moisture, 150 psi CO2 pressure, and die product temperature of 119°C. For the control samples, expansion ratio was maximized at a screw speed of 300 rpm, 22% moisture, and die product temperature of 161°C. Even 27 though the expansion was not as great for the C02-injected samples, these samples typically had a more uniform expansion than the control samples (Figure 3.1.1). For C02-injected wheat starch, Ferdinand et a1. (1990) also reported lower and more uniform expansion for C02-inj ected samples compared to samples conventionally puffed at higher barrel temperatures. Effect of Product Temperature In the study by Ferdinand et a1. (1990), product temperatures in the CO2—injected products were typically maintained below 100°C but did reach as high as 106°C. The researchers suggested that combining two types of structuring methods, such as CO2- injection and steam, had the potential of producing quality expanded products. To investigate this in the present study, barrel temperatures were set to a certain temperature profile but were not further controlled by means such as water-cooling jackets. This method allowed steam to combine with C02 gas as expanding agents. Table 3.1.1 shows the effect of die product temperature on expansion ratio. In most cases, expansion ratio between replicate samples increased with small increases in product temperatures at the die. For example, at 22% moisture, 300 rpm, and 150 psi CO2 pressure, the expansion ratios were 1.8 and 1.4 for product temperatures of 112.7°C and 104°C, respectively. In addition, when moisture content inside the barrel decreased from 25 to 22%, product temperature and expansion ratio both increased (Table 3.1.1). Bhattacharya and Hanna (1987) also reported increases in product temperature and expansion ratio when cornstarch was extruded at lower moisture contents. In the present study, when product temperatures were below 100° C, average expansion ratio did not exceed 1.4. Table 3.1.1. The Effect of Die Product Temperature on Expansion Ratios for Samples in the CO2 (at Lower Temperatures) and Control (at Higher Temperatures) Screw Configurations Screw % Moisture Screw CO2 Product Expansion Config. Speed Pressure Temperature Ratio'I (rpm) (psi) at Die (°C) Control 22 200 N/A 161.3 2.3 Control 22 200 N/A 154.6 2.1 Control 22 250 N/A 160.2 2.6 Control 22 250 N/A 158.7 2.4 Control 22 300 N/A 161.3 2.9 Control 22 300 N/A 159.6 2.7 C02 22 300 0 107.1 1.3 C02 22 300 0 114.7 2.1 CO2 22 300 150 112.7 1.8 C02 - 22 300 150 ‘ 104 1.4 C02 22 350 0 109.2 2.1 C02 22 350 0 1 12.2 2.3 CO2 22 350 150 115.6 2.2 C02 22 350 150 106.6 1.5 C02 22 400 0 106.6 2.2 C02 22 400 0 108.2 2.4 CO2 22 400 150 118.8 2.4 C02 22 400 150 111.5 2.1 CO2 25 300 0 96.3 1.2 CO2 25 300 0 95.9 1.2 C02 25 300 150 99.5 1.3 C02 25 300 150 100.8 1.5 C02 25 350 . O 93.1 1.3 C02 25 350 0 101.2 1.3 CO2 25 350 150 101.0 1.5 C02 25 350 150 103.1 1.6 CO2 25 400 0 95.2 1.3 CO2 25 400 0 98.8 1.5 CO2 25 400 150 102.5 1.6 CO2 25 400 150 101.3 1.8 “ Average based on fifteen measurements. 29 At 25% moisture without C02-inj ection, die product temperatures did not exceed 100°C (Table 3.1.1). When CO2 was injected, product temperatures at the die were slightly over 100°C, indicating that water vapor flash-off, along with C02-injection, contributed to product expansion. At 22% moisture with and without CO2, product temperatures at the die also reached higher than 100°C. At this moisture content, expansion ratios were sometimes higher when no CO2 was injected. Ferdinand et a1. (1990) found similar results, in that when C02 pressure varied from O, 175, 263, and 365 psi at 17% moisture and 200 rpm screw speed, maximum expansion ratio was obtained when CO2 pressure was zero. In the present study, Figure 3.1.2 compares the expansion ratios between samples injected with 150 psi CO2 and 0 psi CO2, at 400 rpm screw speed, and 22% moisture. Figure 3.1.2. shows similar expansion ratios between samples, in addition to uniform expansion. At screw speeds of only 300 rpm and 400 rpm, t-test results at a=0.05 showed that there was not a significant difference between mean expansion ratios at 0 psi or 150 psi CO2 (Appendix 5, Table A.5.9). The C02 screw configuration, die geometry, and feed rate were used at the higher (control) temperatures at 300 rpm, 22% moisture, and 0 psi C02 to investigate the effect of a large temperature increase on expansion ratio. Expansion ratios were 2.3 and 2.4 at die temperatures of 153.4 °C and 155.8 °C, respectively, compared to expansion ratios of 1.3 and 2.1 at die temperatures of 107.1°C and 114.7 °C in the lower temperature profile (Table 3.1.2 compared to Table 3.1.1). It appears that increasing temperature (die temperature increased from 80°C to 150°C) had only a modest effect on expansion ratio, and a maximum expansion ratio was reached. To achieve higher expansion ratios, variables other than temperature must be changed, such as screw configuration and feed 30 rate. For example. the expansion ratios for control samples at the higher feed rate of 5.2 kg/hr reached 2.9 (Table 3.1.1). 0 cm1 2 3 4 5 ' ' ’ ’ l ' l * iilHii if“ tliiti-iiiiliii.iii‘IiiiM ’ilwilii iili Figure 3.1.2. Expansion Ratio Comparison for CO2 samples at 22% Moisture and 400 rpm Screw Speed With or Without C02-Injection. Table 3.1.2. The Effect of Die Product Temperature on Expansion Ratios for Samples in the CO2 Screw Configuration at Higher Temperatures Screw % Moisture Screw Speed C02 Product Expansion Config. (rpm) Pressure Temperature Ratio‘ (psi) at Die (°C) C02 22 300 0 153.4 2.3 CO2 22 300 0 155.8 2.4 " Average based on fifteen measurements. Effect of Screw Speed The relationship between screw speed and average expansion ratio (between replicates) is shown in Figure 3.1.3. The graph shows that expansion ratio increased approximately linearly with increasing screw speed for all samples. For extruded wheat flour, Vainionpaa (1991) found similar trends when barrel temperatures were set to 150°C. In Vainionpaa’s (1991) study, when screw speed increased from 100 to 200 rpm, expansion ratio increased from 1.4 to 1.6. At similar barrel tenrperatures in the present study, expansion ratios were between 2.2 and 2.8 when screw speed increased from 200 to 300 rpm (Figure 3.1.3). 3 2.8 / g 2'6 / 0 Control (High '5 Temperature) g 2.4 ,p ' 0 / :3, 0 0 psr C02, 22 /o E 2.2 g, _ r Morsture E’ a 150 psi C02, 22% g. 2 Moisture Ix: / 00 psi C02, 25% a) 1‘8 ’ fi Moisture a a /a . o 1.6 A I 150 psr C02, 25% E / Moisture 1.4 G O 1.2 G l i i i 100 200 300 400 500 Screw Speed (rpm) Figure 3.1.3. Average Expansion Ratio vs. Screw Speed Using the CO2 (at Lower Temperatures) and Control (at Higher Temperatures) Screw Configurations. 32 Lee et a1. (1999) found a negative relationship between screw speed and expansion ratio when extruded cornstarch was injected with supercritical C02. They attributed this relationship to mean residence time being reduced at higher screw speeds, which most likely caused a reduction in starch gelatinization and expansion. Although starch gelatinization was not measured directly in the present study, it appears that . increasing shear rates with higher screw speeds may have had a greater effect on starch gelatinization than shorter residence times. Relationship Between Screw Speed and SME In addition to theheating elements along the barrel, the dissipation of viScous energy is also a source of heat (Harper, 1978). Figure 3.1.4 shows that increasing average SMEa (average between replicate samples) had a positive effect on average expansion ratio. This trend is similar to the relationship between screw speed and average expansion ratio in Figure 3.1.3. This result suggests that the combination of increasing shear and viscous energy may have contributed to greater starch gelatinization in the dough, thus causing more expansion to occur. This result also indicates that in terms of processing energy costs for a given extrusion set-up, it will cost more to produce a highly expanded product. At 150 psi CO2 for both moisture contents, there was some decrease in expansion ratio at higher SMEs (Figure 3.1.4). Ferdinand et a1. (1990)'also reported that at higher CO2 pressures, expansion ratio increased with increasing SME to a maximum, before gradually decreasing. In the present study, this decrease in expansion ratio with SME at higher C02 pressures may have been the result of inconsistencies in the flow of CO2. ' The average value for dough density was calculated as 1250 kg/m’. 33 Carbon dioxide pressures were monitored with a pressure gauge and pressures stayed fairly constant. However, the actual flow of CO2 into the extruder was not measured directly. Control (High 7 Temperature) 0 psi C02, 22% Moisture 150 psi C02, 22% Moisture 0 psi C02, 25% Moisture 150 psi C02, 25% Moisture l . i . 200.0 300.0 400.0 500.0 600.0 l Average Specific Mechanical Energy (kJ/kg) Average Expansion Ratio Figure 3.1.4. Average Expansion Ratio vs. Average Specific Mechanical Energy Using the CO2 (at Lower Temperatures) and Control (at Higher Temperatures) Screw Configurations. Even though SME is a function of torque, screw speed, and throughput, Altomare and Ghossi (1986) investigated the influence of changing only screw speed on SME in a twin-screw extruder. They reported that at moisture contents of 20 and 25%, SME input increased linearly when screw speed increased fiom 233 to 400 rpm. In the present study, a near linear relationship also existed between SME and screw speed (Figure 3.1.5). 34 The combination of factors influencing SME (increased torque, decreased screw speed, and increased feed rate) caused a net decrease in SME for the control samples compared to the C02 samples. It was reported in Section 1.1.2 that one of the potential advantages to extruding at lower temperatures using C02-injection was reduced energy costs. However, in the present study the opposite result was found and more energy input was required to produce expanded products at lower temperatures than at higher temperatures (Figure 3.1.5). Also, approximately 70% more energy input was required to obtain similar expansion ratios of 2.2 at lower temperatures than at higher temperatures (Figure 3.1.4). A 550.0 g“ A 3 500.0 fit 3 E ‘ 0 Control (High Q A 7 Temperature) ‘5 450-0 5 a 0 psi C02, 22% .3. 9 Moisture .3 400.0 ! ‘ 150 psi C02, 22% 2 Moisture 8 e o o 0 psi CO2, 25% 2 350.0 Moisture $5 I 150 psi C02, 25% .g 3000 ‘ Morsture a. m 3° 250 o g . > ‘3 200.0 . . . 100 200 300 400 Screw Speed (rpm) Figure 3.1.5. Specific Mechanical Energy vs. Screw Speed Using the CO2 (at Lower Temperatures) and Control (at Higher Temperatures) Screw Configurations. 35 3.2. Bulk Density, Water Absorption Index, and Water Solubility Index Bulk Density Average bulk densities for all products extruded at lower temperatures with or without C02 gas were higher than control products puffed at conventional temperatures (Table 3.2.1). Higher bulk densities were likely to occur since the expansion ratios were lower than that of the controls. Ferdinand et a1. (1990) showed that bulk density was lower at a lower moisture content and decreased with increasing screw speed. In the present study, bulk densities in the C02 samples decreased when feed material moisture decreased from 25 to 22%. Also at 22% moisture, bulk density decreased with increasing screw speed (Table 3.2.1). These would be the expected trends for bulk density since expansion ratio increased with increasing screw speed and decreasing moisture content (Figure 3.1.3). No clear trend of bulk density with screw speed was shown at 25% moisture, which may have been due to the low expansion ratios found at this moisture content (Tables 3.1.1 and 3.2.1). WA] and % WSI The products extruded using the C02 screw configuration had lower average WAI than products puffed at higher temperatures (Table 3.2.1). Water absorption at room temperature is related to the extent of starch gelatinization in a extruded product that has been fully cooked. Native starches have zero water absorption capabilities at room temperature due to their compact granular structure, whereas gelatinized starch molecules absorb water at room temperature and swell. As the starch granules swell, WAI increases until a peak viscosity is reached, depending on the extent of starch damage. WAI then 36 decreases with the onset of dextrinization (Colonna et a1. 1989). Since the amount of gelatinized starch molecules was not measured, further analytical tests would be required to determine the extent of starch gelatinization. One method of detennining degree of gelatinization is the Rapid Viscoanalyzer. Table 3.2.1. Average Bulk Density, WA], and % WSI for CO; (at Lower Temperatures) and Control (at Higher Temperatures) Screw Configurations Screw % Screw C02 Average Bulk Average WAI' Average % Config. Moist. Speed Press. Density ' WSI' (mm) (psi) (ran!) Control 22 200 N/A 0.206 E 0.018) 8.95 (:t 0.21) 11.5% (:1: 0.37) Control 22 250 N/A 0.129 (i 0.013 9.18 (a: 0.24) 12.4% (i 1.02) Control 22 300 N/A 0.099 (:1: 0.006) 9.13 (d: 0.16) 13.6% (i 0.59) C02 22 300 0 0.516 (:1: 0.127) 7.00 (2!: 0.67) 12.0 (i 1.68) CD; 22 300 150 0.501 (:t 0.135) 6.66 (:t 0.52) 11.7 (a: 1.4) CO; 22 350 0‘ ' 0.329 (d: 0.040) 7.88 (d: 0.11) 14.0 (i 0.61) C02 22 350 150 0.434 (:1: 0.156) 7.26 (:t 0.19) 11.4 (:1: 1.31) CD; 22 400 0 0.291 (i 0.026) 7.36 (:h 0.34) 13.2 (i 0.44) C02 22 400 150 0.284 (t 0.049) 7.86 (:1: 0.15) 16.6 (:1: 1.47) C02 25 300 0 0.532 (:1: 0.015) 7.92 (:t 2.15 ) 6.71 (i 0.53) CD; 25 300 150 0.605 (:1: 0.125) 5.31 ($0.36) 7.39 (:t 0.29) C02 25 350 0 0.718 (:t 0.066) 7.85 (i 0.57) 7.54 (:t 0.64) C02 25 350 150 0.683 (:t 0.101) 5.62 (:1: 0.11) 8.09 (.4: 0.61) C02 25 400 0 0.689 (i 0.059) 7.42 (d: 0.84) 8.68 (:1: 0.63) C02 25 400 150 0.616 (:1: 0.020) 5.89 ($0.13 9.51 (d: 0.35) ' Average between replicate samples. Standard deviation given in parentheses. Dextrinization is also referred to as starch fragmentation, and causes % WSI to increase. Dextrinization is related to the number of soluble molecules. At a=0.05, t-tests comparing means between high and low-temperature samples at 300 rpm and 22% moisture showed that there was not a significant difference between mean % WSI values (Appendix 5, Table A.5.10). Another t-test comparison, also at a=0.05, was conducted between mean % WSI at 22 and 25% moisture for samples extruded at lower 37 temperatures, 300 rpm screw speed, and 0 psi C02 (Appendix 5, Table A.5.11). As shown in Table A.5.11, % WSI is significantly lower at 25% moisture than compared to 22%. Colonna et a1. (1984) reported for wheat starch that lower water solubilities occur under high moisture and low temperature extrusion conditions- In the present study, C02 samples at 25% moisture had lower SME values than C02 samples at 22% moisture (Figure 3.1.5). These results suggest that the extent of starch gelatinization decreased with decreasing temperature and SME, and increasing moisture content. WA! and % WSI: Eflect of Screw Speed and C02 Within the C02 screw configuration at 25% moisture, % WSI increased with increasing screw speed (Table 3.2.1). When no C02 was injected at 25% moisture only, WAls were higher and % WSIs were lower than the corresponding values for injection of CO; at 25% moisture. At these conditions, extrusion product temperatures were the lowest, indicating that the degree of starch gelatinization may have been reduced. At 22% moisture no apparent trends were found for increasing screw speed or C02 pressure on WAI and % WSI (Table 3.2.1). Bulk Density, WA], and % WSI for Samples Extruded Using the C02 Screw Configuration at Higher Temperatures As stated in Section 3.1, there was a small increase in expansion ratio with a large increase in temperature for the C02 samples at 300 rpm, 22% moisture, and 0 psi C02. The bulk density values in the C02 samples extruded at higher temperatures were lower than the C02 samples extruded at lower temperatures (Table 3.2.1 compared to Table 3.2.2). After drying, the moisture content in the C02 samples at higher temperatures was 38 4.8% compared to 6.7% at the lower temperature profile. This result suggests that for the C02 extrusion conditions, a reduction in bulk density at higher temperatures may have been the result of lower moistures in the final dried product. Average WAI for the C02 samples extruded at higher temperatures was 7.62 (Table 3.2.2). This is higher than the average WAI of 7.00 for the low-temperature C02 samples at 300 rpm (Table 3.2.1). The average % WSI for the high-temperature C02 samples was 14.7 compared to 12.0 in the low-temperature C02 samples (Table 3.2.1 compared to Table 3.2.2). Higher WAI and % WSI values suggest that extent of starch gelatinization was greater with increasing temperature using the C02 extrusion set-up. Table 3.2.2. Average Bulk Density, WA], and % WSI for the CO; Screw Configuration at Higher Temperatures Screw % Screw C01 Average Bulk Average WAI‘ Average % Config. Moist. Speed Press. Density ' (g/ml) WSI‘I 1mm) (psi) co2 22 300 0 0.190 @0007) 7.62 (a 0.37) 14.7% (a: 0.60) ‘ Average between replicate samples. Standard deviation given in parentheses. Statistical Interpretation In Table 3.2.1, average values were the averages between replicate samples (repeated on different days) and in standard deviation was given in parentheses. Standard deviation is a measure of how much the samples varied in physical attributes when extrusion conditions were repeated on a different day. Larger variations in bulk density, WAI, and % WSI appeared to occur only within the C02 screw configuration (Table 3.2.1). In the C02 screw configuration non-uniform mixing was visible in the extruded products only at 25% moiSture and 0 psi C02. In some of the extrudates, there were areas 39 along the strand where residual flour appeared. Therefore, the higher variability in physical characteristics within the C02 screw configuration at 25% moisture may have been due to this occurrence of non-uniform mixing. In Table 3.2.1, there was a higher variability in bulk densities (between replicates) at 22% moisture compared to 25% moisture. In addition, expansion ratios at 22% moisture, were less repeatable than 25% moisture, especially at the lower screw speeds (Table 3.1.1). These greater variations in bulk densities and expansion ratios were probably due to variations in die pressures and die temperatures when extrusion conditions were repeated on different days (Appendix 5, Table A.5.1). 3.3. Mean Residence Time Figures 3.3.1 and 3.3.2 show examples of residence time distribution (RTD) curves, based on one replicate, for the C02 and control screw configurations. At each screw speed, mean residence time was longer in the C02 screw configuration than in the control screw configuration. As stated in Section 2.2, the conveying efficiency of the screw configuration is determined by the geometry of the screw elements. Almost twice as many mixing paddles at a 60° orientation are found in the C02 screw configuration (13 vs. 7, Table 2.1). Since the conveying efficiency of the C02 screw configuration is lowered, the C02 samples spend more time in the extruder barrel than compared to the controls. The RTD curves in the control screw configuration show that as screw speed increased, mean residence time decreased. Altomare and Ghossi ( 1986) also reported that shorter mean residence times resulted when screw speed increased at a constant throughput. The researchers reported that when screw speeds were 150, 317, and 400 40 rpm, mean residence times were 28.4, 22.1, and 20.2 seconds, respectively. In the present study, where a smaller range of screw speeds was used compared to Altomare and Ghossi, there was only a one second difference in mean residence times between screw speeds of 200 and 250 rpm (Figure 3.3.1). When these extrusion conditions were repeated on a different day, the difference in residence times increased to 15 seconds between screw speeds of 200 and 250 rpm (Appendix 5, Table A.5.2), following the expected trend of shorter mean residence times at higher screw speeds. 1‘ Norm allzed Concentration, E(t) 0.0300 0.0250 . 0.0050 0.0000 0.0200 0.0150 . 0.0100 ' + 200 rpm, Mm Residence Time 68 1 seconds 1 i + 250 rpm, Main Residence Time 67 seconds l i \ ‘ - 300 rpm, Mean \ i | Residence Time 58 K ’_7 [ seconds \\ AA 5 4 1 l 0 50 100 150 Tim e (seconds) W 2, +,, W 7 ,,,,Ai W,” W W ,, _,- A AA_,J Figure 3.3.1. RTD Curves for Control (at Higher Temperatures) Screw Configuration at Screw Speeds of 200, 250, and 300 rpm at 22% Moisture. 0.0140 7 7 7 0.0120 77 ~ 7 7 7 77 7 7 r77 7 7 7 0.0100 7 7 -- 7 77777777777777 0.0080 7 7 , 7 7 77 7 777 7 77 7 0.0060 7 7 777 7 7 777777 777 0.0040 7 . 7777 7 .7 0.0020 7777—7777 0.0000 . . . 1 0 50 100 150 200 250 300 Time (seconds) Normalized Concentration, E(t) - 0 ~ 300 rpm, Mean Residence Time 101 seconds 97 seconds 86 seconds Figure 3.3.2. RTD Curves for CO; (at Lower Temperatures) Screw Configuration at Screw Speeds of 300, 350, and 400 rpm at 22% Moisture, and 150 psi C02. Table 3.3.1 lists the mean residence times (between replicates) for the C02 (at lower temperatures) and control (at higher temperatures) screw configurations. For these products, an inverse relationship typically existed between mean residence time and screw speed. At 22% moisture and 0 psi C02 this trend was not followed, which may have been due to the variability in residence times between replicate samples (Table 3.3.1). The higher standard deviations for some of the C02 samples show that it was more difiicult to obtain highly repeatable results than for the control samples. However, the same trends were observed within any one run (Appendix 5, Table A.5.1). 42 Table 3.3.1. Mean Residence Times for CO; (at Lower Temperatures) and Control (at Higher Temperatures) Screw Configurations Screw % Moisture C02 Screw Speed Mean Residence Configuration Pressure (rpm) Time' (seconds) (psi) Control 22 N/A 200 77.08 (:1: 12.1) Control 22 N/A 250 68.29 (:t 1.3) Control 22 N/A 300 57.52 (i 0.2) CO; 22 0 300 129.24 (t 6.2) CO; 22 0 350 136.42 (1 12.2) C0; 22 0 400 116.96 (:1: 23.7) C0; 22 150 300 124.42 (d: 32.3 C0; 22 150 350 114.18 (i 22.9) C0; 22 150 400 \ 100.51 (d: 19.3) C0; 25 0 300 129.73 (i 9.5) CO; 25 0 350 126.98 (2+: 6.7) CO; 25 0 400 118.0 (x 7.7) CO: 25 150 300 135.10 (3: 16.5 C02 25 150 350 128.94 (i 2.6) CO; 25 150 400 103.89 (i 16.4) ‘ Average between replicate samples. Standard deviation given in parentheses. The mean residence time for the C02 screw configuration at higher temperatures was 112.9 :1: 16.2 seconds, compared to the mean residence time of 57.5 i: 0.2 seconds for control samples at 300 rpm (Table 3.3.2 compared to Table 3.3.1). Therefore, even at higher temperatures, the mean residence time for the C02 screw configuration continued to be about twice as long as that of the control screw configuration. Table 3.3.2. Mean Residence Times for the C02 Screw Configuration at Higher Temperatures ' Screw % Moisture C02 Pressure Screw Speed Mean Residence Configuration (psi) (rpm) Time (seconds)' CO; 22 0 300 112.92 (i 16.2) ‘ Average between replicate samples. Standard deviation given in parentheses. 43 3.4. Thiamin Loss Average thiamin losses (based on three measurements) ranged fi'om 10-16% in extrudates puffed using conventional methods (Figure 3.4.1). Guzman-Tello and Cheflel (1987) reported thiamin losses of 30% when the exit product temperature was 160°C for 14% moisture content wheat flour. At 160°C, they also reported that an increase in screw speed caused more thiamin destruction, even though residence time decreased. In the present study, the data also suggests a trend of decreasing thiamin loss with increasing mean residence time (decreasing screw speed) in the controls at higher temperatures (Figure 3.4.1). This trend did not change when samples were extruded using the C02 screw configuration at higher temperatures (Figure 3.4.1). Thiamin losses between replicates were 2.1 and 3.0% at the respective mean residence times of 124.4 and 101.4 seconds (Appendix 5, Tables A.5.3 and A.5.4). It appears that for high-temperature extrusion, thiamin loss increased with increasing screw speed, suggesting that shear effects predominated over thermal effects in thiamin degradation. Compared to the control samples, lower thiamin losses were observed in the C02- injected samples, but only when the moisture content was 22% (Figure 3.4.2). All C02- injected samples at 25% moisture showed greater than 24.5% thiamin loss (Figure 3.4.2). Most of the literature has reported that as feed material moisture content increases, thiamin loss decreases (Maga and Sizer, 1978; Pham and Del Rosario, 1986; Guzman- Tello and Chefiel, 1987). Killeit (1994) suggested that lower thiamin losses at higher moistures were the result of decreased dough viscosities, which reduced the amount of shear and energy input on the product. In the present study, SME values were typically lower for most of the C02 samples at 25% moisture than the C02 samples at 22% 44 moisture (Figure 3.1.5). Unexpectedly, thiamin losses were about 15% greater at 25% _ moisture than the C02 samples at 22% moisture. Hemnann and Tunger (1966) suggested that in the dehydration of foods, changes in moisture contents may significantly affect the concentrations of the constituents and possibly the pH value of the products. Therefore, a higher thiamin loss at 25% vs. 22% moisture in the present work may have been the result of different chemical interactions occurring at the higher moisture content. For example, an increase in feed material moisture content may have resulted in an increase in dough pH, causing more thiamin degradation at 25% moisture. 20 O . 15 77 77 7.777777 777—-7 77 7 7 7 3 . .5 1 O 1 .5 ° 3 l‘ .1: I: '0 ’7—“T—‘ffi—iAnnfiw 77 OControl Screw °\ 1 C nf t'on 3’4: ‘ ‘ o rgurar a 3 i 2 . l 5 "*Ti“fi"_‘”"‘_AIH——'“ ‘ IC02 Screw :‘ l Configtnation I E 1 L _L_ 0 . i . 7 50 70 90 110 130 Mean Residence Time (seconds) L ___ L L L_L_L_LL_L LLLL LLLL_ LL LL L L LL _._1 Figure 3.4.1. Average % Thiamin Loss vs. Mean Residence Time for High Temperature Samples Using the Control and C02 Screw Configurations. 45 40 35 a 30 25 n 20 O 22% Moisture 15 I 25% Moisture Average % Thiamin Loss O O i T T 60 80 100 l 20 l 40 160 Mean Residence Time (seconds) Figure 3.4.2. Average % Thiamin Loss vs. Mean Residence Time for C02-Injected Samples (at Lower Temperatures) at 22 and 25% Moisture. It can be concluded that unlike the high temperature extrudates, it appears that thermal effects had greater influence on thiamin loss than shear effects in the C02- injected samples (Figure 3.4.2 compared to Figure 3.4.1). In Table 3.4.1, thiamin losses and mean residence times for low-temperature samples without C0; are shown. At 22% moisture, % thiamin losses were between 0 and 1.5%, which are lower than the samples at 150 psi C02 and 22% moisture (Table 3.4.1 compared to Figure 3.4.2). At 25% moisture and 0 psi C02, thiamin losses were lower than at 25% moisture and 150 psi C02 (Table 3.4.1 compared to Figure 3.4.2). It appears that for both moisture contents, the addition of C02 gas caused an increase in thiamin loss (Table 3.4.1). In Section 3.1, an 46 increase in die product temperature was observed when CO; was injected, which may have contributed to greater thiamin losses (Table 3.1.1). Table 3.4.1. Average % Thiamin Losses and Mean Residence Times Without CO; at Lower Temperatures. % Moisture Screw Speed Average % Mean residence (rpm) Thiamin Loss' time (seconds) 22 300 0 (:t 2.2) 124.8 22 300 1.5 (:t 0.8) 133.6 22 350 0 (i 2.9) 127.8 22 350 0 (i 0.8) 145.0 22 400 0 (:t 4.1) 100.2 22 400 0 (:1: 1.7) 133.7 25 300 26.4 (:t 1.1) 136.4 25 300 0.5 (i 0.4) 123.0 25 350 16.1 (:t 0.3) 131.7 25 350 1.1 (i 1.5) 122.3 25 400 11.6 (:i: 1.0) 112.6 25 400 0 (2t 1.5) 123.4 3 Average of three measurements. Standard deviation given in parentheses. As shown in Table 3.4. 1, thiamin loss was greater at 25% moisture than compared to 22% moisture. The variability in % thiamin loss among replicates at 25% moisture was also greater than the samples at 22% moisture (Table 3.4.1). Despite the variability, the trend of increasing thiamin loss with increasing mean residence time was typically followed at 25% moisture. As stated above for the control screw configuration, thiamin loss decreased with longer residence times. However, a trend of higher thiamin losses with longer mean residence times for the C02 samples is shown in Figure 3.4.2. In Appendix 5 (Tables A.5.6 and A57), the maximum % CV for % thiamin (w/w) in the feed flour was 6.0% for three l-gram samples. Guzman-Tello and Chefiel (1987) reported the % CV of five 47 30-gram samples to be 3.6% when flour and thiamin were mixed in a high speed mixer for one hour. The % CV for % thiamin in extruded products did not exceed 5%, indicating that the analytical method for thiamin analysis was precise (Tables A.5.6 and A.5.7). 48 Conclusions and Recommendations 4.1 Summary and Conclusions Physical characteristics were investigated in puffed-wheat flour extruded at lower temperatures (near 100°C) using carbon dioxide gas as the expanding agent. Specific mechanical energy inputs, mean residence times and thiamin losses were also investigated. These dependent variables in low-temperature extrusion were compared to puffed-wheat extruded at conventional puffing temperatures (160°C). The current work shows that using the screw configuration listed in Chapter 2 (“C02-Injection”, Table 2.1) at a barrel temperature profile of 40/40/50/70/80°C, with C02 pressures of 0 and 150 psi, die geometry of 3 mm diameter x 6mm length, feed material moisture content of 22%, and using screw speeds and feed rates resulting in SMEs between 426 and 572 kJ/kg, and mean residence times in the range of 87 to 145 seconds, extruded wheat can be expanded (expansion ratios > 2.0) at lower temperatures; Even though the samples extruded at lower temperatures had more uniform expansion, the expansion ratios were still lower than those of the high-temperature control samples. The expansion ratios for C02- ' injected samples at 22% moisture (1.4—2.4) were closer to the control than the C02- injected samples at 25% moisture (1.2-1.8). When no C02 was injected, the expansion ratios were 1.3-2.4 and 1.2-1.5, for 22 and 25% moisture, respectively. Expansion ratio was higher with increasing product temperature, screw speed, and energy input. This trend of increasing expansion with increasing energy input indicates that it will cost more to process a more highly expanded product. At 22% moisture, even though barrel temperatures were set below 100°C, product temperatures in the C02-injected samples did reach up to 119° C, indicating that 49 expansion was due to the combination of water vapor flash-off and C02-injection. When no C02 was injected, product temperatures reached as high as 115°C. At 22% moisture, for screw speeds of 300 and 400 rpm, t-Test (0L=0.05) results showed that C02-injection did not have a significant effect on expansion ratio. Consistent with the expansion ratio results, bulk densities were higher in the samples extruded at lower temperatures compared to the high-temperature control samples. Bulk densities in the C02 samples decreased when feed material moisture decreased from 25 to 22%. At 22% moisture only, bulk density decreased with increasing screw speed. In comparison to the control, all samples extruded at the lower temperature profile had lower WAIs. The % WSI values decreased with decreasing temperature and SME, and increasing moisture content. Lower % WSI values in the C02 samples at 25% moisture may be an indication that extent of starch gelatinization was lower than compared to the other samples. A Rapid Viscoanalyzer (RVA) is one way to measure the extent of starch gelatinization. A texture-measuring device could also be used to see how degree of starch gelatinization correlated with product texture. To obtain a better assessment of “bowl-life” in the low-temperature samples, additional texture measurements could be taken after soaking the extruded products in water. Typically, for both screw configurations (C02 and control) mean residence time decreased as screw speed increased. The only condition that did not follow this trend was at 22% moisture and 0 psi C02, and may have been due to the variability in residence times when extrusion conditions were repeated. The C02 screw configuration had mean 50 residence times that were approximately twice as long as the high-temperature screw configuration. Percent thiamin loss for conventionally puffed products was 10-16% and thiamin loss values decreased with decreasing screw speed. The opposite trend was found for C02-injected products, where thiamin loss increased with decreasing screw speed. These opposite trends suggest that for high temperature extrusion, mechanical effects predominate in thiamin degradation, and for lower temperature extrusion with C02- injection, thermal effects predominate. In the C02-injected products at 22% barrel moisture, thiamin loss was 3-11%. At 25% barrel moisture, thiamin loss was 24-34%. Processing variables, such as SME and product temperatures, were lower at this moisture content than at 22% moisture. Higher losses at 25% moisture may have been the result of different chemical interactions occuning at the higher moisture content, such as an increase in dough pH at 25% moisture. At 22 and 25% moisture, product temperatures at the die were lower (up to 10°C) when no CO; was injected compared to when C02 was injected. Thus, suggesting that higher temperatures caused increased thiamin loss. Thiamin losses at 22% moisture without C02 were between 0 and 1.5%. Thiamin losses ranged from 0 to 26.4% at 25% moisture without C02, and more variability occurred between these replicate samples than compared to 22% moisture. Despite the variability, trends of increasing thiamin loss with increasing mean residence time were typically followed. When samples were extruded using the C02 screw configuration at the higher temperature profile, feed rate of 3.6 kg/hr, 3 mm diameter x 6 mm length, 300 rpm screw speed, and 22% moisture content, expansion ratio was approximately 2.4. Between 51 replicates, average percent thiamin loss at these conditions was 2.5% when product temperatures reached as high as 156°C. The mean residence time using the C02 screw configuration at higher temperatures was also approximately twice as long as in the control screw configuration at 300 rpm screw speed. These results indicate that when a different extrusion set-up is used at higher temperatures, shear effects still predominate in thiamin degradation over thermal effects. Thus, suggesting that changing screw configuration, feed rate, and die geometry in higher temperature extrusion has the potential of decreasing thiamin loss in extruded puffed-wheat. 4.2. Recommendations for Future Research It is important to note that the results summarized in Section 4.1 were found using a lab-scale extruder, and the range of extrusion conditions described in Section 2.2.1, 2.2.2, and 2.8. Because of these limitations, it is difficult to determine whether or not there is an advantage to using C02-injection when extruding wheat at lower temperatures. Therefore, the following topics are recommended for future research: 1. Further optimize C02 injection set-up by changing screw configuration and point of C02 injection. Determine how these changes will impact residence time, retention of thiamin or other heat- and/or shear-sensitive additives (i.e. anthocyanins or beta-carotene), and physical attributes. Also add a flow meter at point of C02 injection to precisely measure amount of C02 gas entering extruder. 2. Investigate the effect of screw configuration, feed rate, and die geometry on the loss of heat- and/or shear-sensitive additives at higher temperatures. 52 Select a set of screw configurations with varying levels of conveying efficiency. At a constant initial concentration, investigate only how changes in those process variables affect the loss of the additives. Repeat the experimental design in Table 2.3 and use equipment, such as RVA, to determine degree of starch gelatinization. Investigate the relationship between degree of starch gelatinization on expansion ratio, bulk density, WAI, and % WSI. Also use a texture-measuring device to correlate texture with degree of starch gelatinization. Compare all results in the low-temperature products to hi gh-temperature controls. Set-up an experimental design with a wide range of moisture contents and C02 pressures. Keeping initial concentration of heat- and/or shear-labile additives and other processing variables constant, investigate only the effects of C02 pressure and moisture content on degradation of the additives. 53 Appendices 54 Appendix 1 55 Appendix 1. Mixing Method, Flour and 0.3% Thiamin Raw Materials: 4.5 g Thiamin Hydrochloride (5% Moisture) + 1495.5 g Wheat Flour (13% Moisture) Total Batch Weight: 1500.0g Equipment: 0Hobart Corporation (Troy, Ohio) 5-quart all-purpose mixer with mixing bowl. OPatterson-Kelley twin-shell dry blender. 1. Weigh out 4.5 g thiamin hydrochloride in weighing container. 2. Weigh out 500 g wheat flour in mixing bowl. 3. To 500 g flour add V2 of the thiamin. Place the flat mixing paddle on mixer head and mix on low for 1 minute. Add remaining thiamin and mix on low for 1 minute. 4. In a second mixing bowl weigh out 500 g of wheat flour. Do not tare scale. Add flour/thiamin mixture on top of 500 g of wheat flour, scale should read approximately 1004.5 g. Add more wheat flour until scale reads 1500 g. 5. Take 1500 g of flour/thiamin mixture and spoon it into twin shell dry blender. Set mixing time for 40 minutes and begin mixing (inner bar speed measured by manual tachometer = 1400 rpm). Scrape down blender twice during mixing. During scrape down, also take about 200g out from bottom, add it back to the top of the mixer, and resume mixing. 6. Empty contents of mixer and put aside in a sealed container until needed. Table A.l. Physcial and Chemical Properties of Soft White Wheat Flour (Star of the West Milling Co.)* Moisture 12.75-13.5% AOAC 925.10 Protein 7.5-8.1% AACC 96-12 Ash 0.42-0.47 AOAC 936.07 pH 5.9-6.1 (unbleached) AOAC 943.02 Falling Number 250 seconds AOAC 976.13 * As stated in the manufacturer specification sheet. 56 Appendix 2 57 Appendix 2. Modifications to AOAC Thiamin Analysis Method Study 1. Investigation of Salt Amount on Thiochrome Yield Methods: Varying amounts of salt were added to 7 different 2.5 ml standard aliquots (concentration 15.6 ug/ml) and the oxidation step then proceeded normally with the addition of oxidizing reagent and isobutanol (Appendix 4). Results: Fluorescence 1 1.5 2 2.5 3 Amount Salt (9) I Figure A.2.1. The Effect of Salt Mass on Thiochrome Yield. Conclusion: The amount of salt added in the oxidation step does affect thiochrome yield (directly proportional to fluorescence). It appears that when salt is present in the final separation of aqueous and isobutyl layers, it decreases thiochrome yield. 58 Study 2. Determination of the optimal amount of potassium ferricyanide to oxidize thiamin to thiochrome Method: Varying amounts of 0.4069% potassium ferricyanide solution were added to 7 different 2.5 m1 aliquots of standard solution (concentration 15.6 ug/ml) and 15% NaOH. The reaction then proceeded as normal with the addition of isobutanol (Appendix 4). Results: 395.00 .. . . . . . ,7. . .. “7...,3L}: 393.00 " ' ' ‘ " ' ’ ' " f 391.00 “‘ ‘ ‘ I I 3 389.00 b ; ,. ‘ . A . . 5 387.00 ' ' ' I " . . .. . . . 3 385.00 4—1 ‘ . . . ‘ L- : .. 5' . . . - , ,.- ‘,'--" ;- .. .. .1 -3 ‘“ 381.00 , 1 .. 4 Jo r ' I a 379.00 p 1 p . ‘ ‘ 375.00 A, ' . ‘ . ' ; 9 .L- 0.00 0.20 0.40 0.60 0.80 1.00 Volume 0.004096 glmL Potassium Ferricyanide Solution 1le Figure A.2.2. The Effect of Potassium Ferricyanide Amount on Thiochrome Yield. Conclusion: The optimal amount of 0.4069% potassium fenicyanide is between 0.25 and 0.38 ml’s, for converting thiamin to thiochrome. This amount is 1/1000 of the amount used in the AOAC method. Therefore, for a 2.5 mL aliquot of standard solution at 100 ug/mL, approximately 0.001024 to 0.001556 grams of potassium ferricyanide are needed to optimally oxidize thiamin to thiochrome. 59 Appendix 3 60 Appendix 3. Standard Curves for Fluorescence vs. Thiamin Concentration (pg/ml) 600.00 500.00 100.00 0.00 0 0.5 l 1.5 2 2.5 3 3.5 4 Thiamin Concentration (rig/ml) Figure A.3. Standard Curves (Fluorescence vs. Thiamin Concentration) Constructed in Duplicate on Two Difierent Days. 61 Appendix 4 Appendix 4. Thiamin Analysis Method Introduction: This method was developed from the 1995 AOAC “Official Method 953.17 Thiamin (Vitamin B.) in Grain Products, Fluorometric (Rapid)” and modified for the analysis of extruded wheat flour that has been fortified with 0.3% (w/w) thiamin. The volumes and masses of reagents used in the method were optimized for the specific product under investigation. Using the extruded wheat product and the quantities listed, the maximum thiochrome concentration being analyzed by the fluorometer is 2.6 ug/mL. The standard is prepared at 3.6 ug/mL to cover the entire concentration range needed. Reagents: Sodium hydroxide solution - 15%. Dissolve 15g NaOH in H20 to make 100 mL. Oxidizing Stock Solution. - 0.2796% Dissolve 0.2796g K3Fe(CN)6 in 15% NaOH Solution to make 100 mL. Prepare solution on day it is used. Oxidizing Working Solution - 0.01398% Take a 10 mL aliquot of the oxidizing stock solution and dilute it to 200 mL with 15% NaOH. Prepare solution on day it is used. Isobutyl alcohol (2-ntethyl-Ipropanol )- Use HPLC grade. Quinine sulfate stock solution — Use quinine sulfate solution to govern reproducibility of fluorometer. Prepare stock solution by dissolving 10 mg quinine sulfate in 0.1N H2S04 to make 1 L. Store in light-resistant containers. Quinine sulfate standard solution — Dilute 1 volume quinine sulfate stock solution with 39 volumes 0.1N H2804. (Solution fluoresces to ca same degree as does isobutanol extract of thiochrome obtained from 1 ug thiamine-HCL.). Store solution in light- resistant container. Thiamin hydrochloride standard solutions — (1) Stock solution. -- lOOug/mL. Accurately weight 50-60 mg USP Thiamin Hydrochloride Reference Standard that has been dried to constant weight over P205 in desiccator (Reference standard is hygroscopic; avoid absorption of moisture.) Dissolve in 20% alcohol adjusted to pH 3.5 —- 4.3 with HCl, and dilute to 500 mL with the acidified alcohol. Add enough additional acidified alcohol to make concentration exactly 100 pg thiamine-HCI / mL. Store at ca 10°C in glass-stoppered, light-resistant bottle. 63 Preparation of Standard Solution: Dilute 15.6 mL thiamin-HCl stock solution to 100 mL with ca 0.1N HCl (1 mL = 0.2 pg thiamin-HCI). Note this concentration as (Std to“). Designate this as working standard solution. If NaCl is to be added to sample for extraction, add NaCl to working standard solution, before final dilution, to give final concentration of ca 5% [weight / volume]. (add about 5.0 g NaCl to give correct concentration) Extraction: 95 -1 00 °C Digestion Add ca 1 gram of sample and ca 10 grams salt in 250 mL centrifuge tube. (Addition of about 10 g NaCl to give final concentration of ca 5% [weight / volume] aids in subsequent separation of sample solution. Note amount of sample used (Snip um) and the amount of salt used (Salaam). Thoroughly mix flour and salt with stirring rod before adding 0.1N HCl.) Add 200 mL 0.1 N HCl solution to flour and salt in 250 mL centrifuge tube, using part of acid to wash down sides of vessel. Note amount of acid added (nga). Shake vessel to thoroughly mix contents. Place vessel in H20 bath previously heated to 95-100°C. With caps on vessels, stir at frequent intervals to keep solids in suspension during thickening stage (5-8 min) for total heating time of 30 minutes. A shaking water bath is an acceptable way to heat and shake the tubes. Afier hydrolysis has proceeded, place a drop of solution on a spot plate and test with thymol blue indicator. Solution should be distinctly red (pH 1.0-1.2). If not (indicating the presence of basic substances in sample), add ca 1N HCl in 1.0 mL portions until desired acidity is reached. Note volume of 1N acid required to supplement the 0.1N acid and REPEAT DIGESTION with new sample weight and necessary mixture of 1N and 0.1 N acids. Cool tubes to room temperature. [Note: This pH test only needs to-be done initially, and checked periodically thereafter] Centrifuge tubes at 5,000 rpm (4068 x g) for 10 minutes. Oxidation: [Note: In the paragraphs that follow, the specific order in which reagents were added in the oxidation step is described. This order was tested in previous studies and was shown to produce the highest thiochrome yield when compared to adding the reagents in a different sequence] Standard Solutions: Using this method, standard curves do not have to be run on every day that a sample is analyzed. Standard curves should be checked periodically between days of analysis to 64 confirm low variability (<5%) between slopes. Once low variability is established, average y-intercept and slope values are to be used in the final calculation. The standard curve is done at four concentrations in'duplicate, making for 8 oxidized tubes and 8 un-oxidized tubes. The setup for one set of tubes is shown below. Table A.4. Reagent Amounts for the Construction of a Standard Curve (Fluorescence vs. Thiamin Concentration) Tube Oxidized/ Volume Ending Thiamin Volume Volume # Non-oxidized Standard Concentration Oxidizing 15% NaOH Added (pg/mL) Solution Solution (mL) (mL) (mL) 1 Oxidized 3.00 3.60 12.00 2 Oxidized 2.25 2.70 9.00 3 Oxidized 1.55 1.86 6.20 4 Oxidized 0.80 0.96 3.20 5 Non-Oxidized 3.00 3.60 12.00 6 Non-Oxidized 2.25 2.70 9.00 7 Non-Oxidized 1 .55 1.86 6.20 8 Non-Oxidized 0.80 0.96 3.20 Oxidized Tubes To 1 ca 40 mL tubes (or reaction vessels) add necessary amount of working standard solution (amounts indicated in Table 4A.1). UNIFORM TECHNIQUE MUST BE USED. Protect solution from sunlight, which destroys thiochrome. Use pipet that delivers 10 mL in 1-2 sec for addition of oxidizing reagent. Place tip of pipet containing oxidizing reagent in neck of tube and hold it so that the stream of solution doesn’t hit side of tube. Gently swirl tube to produce rotary motion in liquid and immediately add necessary amount of oxidizing reagent (amounts indicated in Table 4A.1). IMMEDIATELY add 13 mL isobutanol, stopper, and shake for at least 15 sec. Afier isobutanol has been added to all tubes, shake again for ca 2 minutes (tubes may be placed in a shaker box for this additional shaking). Centrifuge tubes at ca 10,000 rpm (11,950 x g) for 5 minutes. Pipet ca 3.3 mL isobutanol extract (upper layer) from each tube into sample cuvette for thiochrome fluorescence measurement. Non-Oxidized Tubes To 1 ca 40 mL tubes (or reaction vessels) add necessary amount of working standard solution (amounts indicated in Table 4A.1). UNIFORM TECHNIQUE MUST BE USED. Protect solution from sunlight which destroys thiochrome. Use pipet that delivers 10 mL in 1-2 sec for addition of 15% NaOH solution. Place tip of pipet containing 15% NaOH solution in neck of tube and hold it so that the stream of solution 65 doesn’t hit side of tube. Gently swirl tube to produce rotary motion in liquid and immediately add necessary amount of 15% NaOH solution (amounts indicated in Table 4.A.1). Remove pipet and swirl again to ensure adequate mixing. IMMEDIATELY add 13 mL isobutanol, stopper, and shake for at least 15 sec. After isobutanol has been added to all tubes, shake again for ca 2 minutes (tubes may be placed in a shaker box for this additional shaking). Centrifirge tubes at ca 10,000 rpm (11,950 x g) for 5 minutes. Pipet ca 3.3 mL isobutanol extract (upper layer) from each tube in to cuvette for thiochrome fluorescence measurement. Sample Solutions Oxidized Tubes To 1 ca 40 mL tubes (or reaction vessels) add 2.5 mL sample solution from the 250 mL centrifuge tube. UNIFORM TECHNIQUE MUST BE USED. Protect solution from sunlight, which destroys thiochrome. Use pipet that delivers 10 mL in 1-2 see for addition of oxidizing reagent. Place tip of pipet containing oxidizing reagent in neck of tube and hold it so that the stream of solution doesn’t hit side of tube. Gently swirl tube to produce rotary motion in liquid and immediately add 10 mL oxidizing reagent. Remove pipet and swirl again to ensure adequate mixing. IMMEDIATELY add 13 mL isobutanol, stopper, and shake for at least 15 sec. Non-Oxidized Tubes ' To 1 ca 40 mL tubes (or reaction vessels) add 2.5 mL sample solution from the 250 mL centrifuge tube. UNIFORM TECHNIQUE MUST BE USED. Protect solution from sunlight which destroys thiochrome. Use pipet that delivers 10 mL in 1-2 sec for addition of 15 % NaOH. Place tip of pipet containing 15 % NaOH in neck of tube and hold it so that the stream of solution doesn’t hit side of tube. Gently swirl tube to produce rotary motion in liquid and immediately add 10 mL 15% NaOH. Remove pipet and swirl again to ensure adequate mixing. IMMEDIATELY add 13 mL isobutanol, stopper, and shake for at least 15 see. Note amount of sample solution added to reaction vessel as ml. mp .uqm Note amount of isobutanol added to reaction vessel as mL mp “ohm“, Afier isobutanol has been added to all tubes, shake again for ca 2 minutes (tubes may be placed in a shaker box for this additional shaking). Centrifuge tubes at ca 10,000 rpm (11,950 x g) for 5 minutes. Pipet ca 3.3 mL isobutanol extract (upper layer) fi'om each tube in to cuvette for thiochrome fluorescence measurement. 66 Thiochrome Fluorescence Measurement: Standardizing the F luorom eter To standardize the fluorometer, set excitation at 343 nm and emission at 459 nm, sensitivity at 10, and selector at 1. With nothing in fluorometer, use zeroing knob to zero machine. Put quinine standard in sample cell, and in fluorometer. Using variable knob, set machine to 10.00. Measuring Fluorescence Construction of Standard Curve After standardizing, set excitation at 373 nm and emission at 410 nm. Change sensitivity to 1/10 and using variable knob, set machine to 0.00. Measure fluorescence (S) of 3.3 mL of extract from oxidized assay standard solution. Change sensitivity to 10 and measure fluorescence (d) of 3.3 mL extract from assay standard solution which has been treated with 15% NaOH (standard blank). After fluorometric readings have been recorded, calculate (S-d) for each standard solution of known concentration. Using Microsofi Excel, plot fluorometric readings (S-d) vs. thiamin concentrations (uglmL) and derive a linear equation from the plotted data. Measuring Sample Fluorescence Afier standardizing, set excitation at 373 nm and emission at 410 nm. Change sensitivity to 1/ 10 and using variable knob, set machine to 0.00. Measure fluorescence (I) of 3.3 mL of isobutanol extract from oxidized assay sample solution. Change sensitivity to 10 and measure fluorescence (d) of 3.3 mL extract from assay standard solution which has been treated with 15% NaOH (sample blank). 67 Calculation for % thiamin in sample (pg thiamin/g sample): 1. Use standard curve to solve for thiamin conc. (pg/mu: PC = 13 m Where, y = fluorometric reading (oxidized — blank) 2 = average y-intercept PC = thiamin concentration in fluorometric cell, pg/mL m = average slope 2. % thiamin calculation: mL . '° FCx (%+m+mLm) x 1 x ’"m"s°b""'"°‘ x 10 g x 1 x100% FD smp. aliquot 1 #g Smpm Where, SM = salt mass (grams) SD = salt density (g/ml) FM = flour mass (grams) FD = flour density (g/ml) 3. Percent thiamin calculation given the average slope is 119.4 and average y- intercept is 51.2, and the measured salt and flour densities are 2.17 g/mL and 0.48 g/mL respectively: -6 2.62%x mg 7 094g +200mL x—l-xflxflx 1 x100%=0.2997% g T g 2.5 1 0.94 2.17 m 0.48 /mL #2 g 0 % thiamin (dry basis) for flour (13.5 % moisture, w/w) = (212—91932? = 0.3465% 68 Appendix 5 69 Table A.5.1. Processing Data and Expansion Ratio, and Bulk Density Values for Appendix 5. Extrusion Data CO; (Lower Temperature) and Control (Higher Temperature) Extrusion Conditions Screw '/. Screw C01 Melt Temp. Product °/o Die Exp. Bulk Conflg. Moist. Speed (psi) (°C) Temp. Torq. Press. Ratio‘ Dens. (rpm) at Die (ps1) (gllul)b (”Cl Control 22 200 N/A 41/75/106/ 157.8 65 800 2.3 0.222 138/165 ($0.44) ($0.006) Control 22 200 N/A 44/79/102/ 154.6 47.5 620 2.1 0.190 1 36/164 ($0.59) ($0.008) Control 22 250 N/A 40/76/105/ 160.2 56.0 720 2.6 0.141 140/164 ($0.43) ($0.007 Control 22 250 N/A 41/77/104/ 158.7 55.7 720 2.4 0.118 138/164 ($0.53) ($0.006) Control 22 300 N/A 41/76’127/ 161.3 55.0 700 2.9 0.093 143/164 ($0.45) ($0.002) Control 22 300 N/A 41/77/104/ 159.6 53.0 680 2.7 0.104 142/ 1 63 $0.42) ($0.000) CO; 22 300 0 41/61/71/89/96 107.1 32.5 600 1.3 0.401 ($0.02) ($0.000) CO; 22 300 0 41/62/71/91/97 114.7 35.9 820 2.1 0.604 ($0.23) ($0.005) CO; 22 300 150 41/64/72’85/97 1 12.7 44.0 940 1.8 0.626 ($0.28) ($0.004 C02 22 300 150 41/59/69/86/93 104 32.3 600 1.4 0.832 ($0.31) ($0.01) C02 22 350 0 41/60/71/92/97 109.2 34.5 690 2.1 0.592 ($0. 19) ($0.006 CO; 22 350 0 41/60/70/91/96 1 12.2 35.6 790 2.3 0.410 ($0.10) ($0.002 C02 22 350 150 41/63/72/86/97 1 15.6 34.0 700 2.2 0.452 ($0.23) ($0.004) CO; 22 350 150 41/60/69/85/93 106.6 31.8 620 1.5 0.742 ($0.14) ($01E) CO; 22 400 0 41/56/65/82/89 106.6 30.0 610 2.2 0.482 ($0.17) ($0.005 CO; 22 400 0 41/58/67/84/91 108.2 31.2 760 2.4 0.429 ($0.11) ($0.004) C02 22 400 150 41/65/75/88/98 1 18.8 36.0 700 2.4 0.484 ($0. 19) ($0.005) C02 22 400 150 41/59/69/87/95 11 1.5 30.9 600 2.1 0.479 ($0.20) ($0.008) C02 25 300 0 42/63/76/84/85 96.3 32.5 335 1.2 0.845 ($0.10) ($0.008) CO; 25 300 0 41/61/71/83/89 95.9 31.5 400 1.2 0.787 ($0.02) ($0.12) C02 25 300 150 41/63/74/87/91 99.5 34.4 530 1.3 0.739 ($0.14) ($0.005) CO; 25 300 150 41/69/85/94x9l 100.8 36.2 300 1.5 0.860 ($0.15) ($0.014) CO; 25 350 0 41/62/75/82/84 93.1 32.0 340 1.3 0.830 ($0.12) ($0.005) C02 25 350 0 41/61/71/86/92 101.2 34.5 540 1.3 0.850 ($0.03) ($0.017) CO; 25 350 150 41/61/74/87/92 101.0 33.1 560 1.5 0.880 ($0.13) ($0.004) C02 25 350 150 41/70/86/95/92 103.1 34.5 330 1.6 0.886 ($0.39) ($0.008) 70 Table A.5.l. Processing Data and Expansion Ratio, and Bulk Density Values for C02 (Lower Temperature) and Control (Higher Temperature) Extrusion Conditions (cont’d) Screw % Screw CO; Melt Temp. Product '/o Die Exp. Bulk Config. Moist. Speed (psi) (°C) Temp. Torq. Press. Ratio‘ Dena. (rpm) at Die (psi) (g/ml)" 00 co; 25 400 0 41/57/68/77/85 95.2 29.3 290 1.3 0.816 ($0.13) ($0.003 C02 25 400 0 44/59/69/84/89 98.8 34.1 560 1.5 0.817 ($0.1 Q ($0.006) C0. 25 400 150 44/59/69/87/92 102.5 34.0 600 1.6 0.759 (9.030) ($0.009) C02 25 400 150 41/60/84/89/91 101.3 32.5 300 1.8 0.816 ($0.50) ($0.020) ' Average values based on two measurements. Standard deviation given in parentheses. b Average value based on three measurements. Standard deviation given in parentheses. ° Average value based on fifteen measurements. Standard deviation given in parentheses. 71 Table A.5.2. Processing Data and WA], % WSI, SME and Mean Residence Time Values for CO; (Lower Temperature) and Control (Higher Temperature) Extrusion Conditions Screw '/o Screw C0; Melt Temp. Product % Die WAI' % SME Mean Config. Moist. Speed (psi) (°C) Temp. Torq. Press. WSI‘ (Id/kg) Res. (rpm) at Die (ps1) Time (°C) (’1 Control 22 200 N/A 41/75/106/ 157.8 65 800 9.26 1 1.2 342.2 68.5 138/165 ($0.06) ($0.06) Control 22 200 N/A 44/79/102/ 154.6 47.5 620 9.00 1 1.8 249.6 85.6 136/164 ($0.28) ($0.8m Control 22 250 N/A 40/76/105/ 160.2 56.0 720 9.36 13.2 369.3 67.4 140/164 ($0.49) ($0.99) Control 22 250 N/A 41/77/104/ 158.7 55.7 720 9.00 1 1.6 367.0 69.2 138064 ($0.18) $1.1) Control 22 300 N/A 41/76/127/ 161.3 55.0 700 9.07 13.9 436.1 57.7 143/164 ($0.09) ($0.57) Conmol 22 300 N/A 41/77/104/ 159.6 53.0 680 8.82 13.3 419.9 57.3 142/163 (ii-l) ($1.3) CO; 22 300 0 41/61/71/89/96 107.1 32.5 600 6.41 10.5 386.9 124.8 ($0.03) ($0.40) C02 22 300 0 41/62/71/91/97 114.7 35.9 820 7.58 13.4 426.2 133.6 ($0.30) ($0.26) C02 22 300 150 41/64/72/85/97 112.7 44.0 940 7.12 12.9 522.4 101.6 ($0.1 1) ($0.90) CO; 22 300 150 41/59/69/86/93 104 32.3 600 6.21 10.5 383.9 147.3 ($0.12) ($0.35) CO; 22 350 0 41/60/71/92/97 109.2 34.5 690 7.97 14.5 479.4 127.8 ($0.08) ($0.07) C02 22 350 0 41/60/70/91/96 112.2 35.6 790 7.79 13.5 494.0 145.0 ($0.48) ($0.55) C02 22 350 150 41/63/72/86/97 115.6 34.0 700 7.42 12.6 471.9 98.0 ($0.03) ($0.05) CO; 22 350 150 41/60/69/85/93 106.6 31.8 620 7.10 10.3 441.4 130.4 ($0.24) ($0.31) CO; 22 400 0 41/56/65/82/89 106.6 30.0 610 7.65 13.6 476.9 100.2 ($0.25) ($0.68) CO; 22 400 0 41/58/67/84/91 108.2 31.2 760 7.07 12.9 495.0 133.7 ($0.14) ($0.44) CD; 22 400 150 41/65/75/88/98 1 18.8 36.0 700 7.73 17.9 571.9 86.9 ($0.04) ($1.0) C02 22 400 150 41/59/69/87/95 1 11.5 30.9 600 8.00 15.3 490.8 114.1 ($0.22) ($0.15) C02 25 300 0 42/63/76/84/85 96.3 32.5 335 9.78 7.0 374.3 136.4 ($0.43) ($0.45) C02 25 300 0 41/61/71/83/89 95.9 31.5 400 6.07 6.4 362.4 123.0 ($0.06) ($0.52) C02 25 300 150 41/63/74/87/91 99.5 34.4 530 5.00 7.5 396.0 123.4 ($0.24) ($0.38) C02 25 300 150 41/69/85/94/91 100.8 36.2 300 5.62 7.3 417.4 146.8 ($0.09) ($0.47) CO; 25 350 0 41/62/75/82/84 93.1 32.0 340 8.31 8.2 430.3 131.7 ($0.55) ($0.03) CO; 25 350 0 41/61/71/86/92 101.2 34.5 540 7.40 7.0 463.0 122.3 ($0.07) ($0.48) CO; 25 350 150 41/61/74/87/92 101.0 33.1 560 5.52 8.6 444.7 127.1 ($0.03) ($0.13) CO; 25 350 150 41/70/86/95/92 103.1 34.5 330 5.72 7.6 464.1 130.8 ($0.02) ($0.21) 72 Table A.5.2. Processing Data and WAI, % WSI, SME and Mean Residence Time Values for CO; (Lower Temperature) and Control (Higher Temperature) Extrusion Conditions (cont’d) Screw % Screw C0; Melt Temp. Product % Die WAI' % SME Mean Contig. Moist. Speed (psi) (°C) Temp. Torq. Press. WSI‘ (Id/kg) Res. (rpm) at Die (psi) Time (°C) (3) C02 25 400 0 41/57/68/77/85 95.2 29.3 290 6.72 8.7 450.6 112.6 ($0.05) ($0.21) C02 25 400 0 44/59/69/84/89 98.8 34.1 560 8.12 8.9 523.2 123.4 ($0.0Q ($1.5) C02 25 400 150 44/59/69/87/92 102.5 34.0 600 5 .98 9.8 522.3 92.3 ($0.01) ($0.11) C02 25 400 150 41/60/84/89/91 101.3 32.5 300 5.80 9.2 500.0 115.5 ($0.01) ($0.35) ' Average values based on two measurements. Standard deviation given in parentheses. b Average value based on three measurements. Standard deviation given in parentheses. ‘ Average value based on fifteen measurements. Standard deviation given in parentheses. 73 Table A.5.3. Processing Data and Expansion Ratios, Bulk Densities, WA], % WSI, SME, and Mean Residence Time Values for CO; Extrusion at Higher Temperatures, 22% Moisture, Screw Speed of 300 rpm, and 0 psi C02 Melt Temp. Product ./0 Die Exp. Bulk Dena. WAI' % SME Mean (‘0 Temp. Torque Press. Ratio‘ (g/ml)h WSI' (Id/kg) Res. at Die (ps1) Time (a) (°C) 44/75/102/ 153.4 24.9 300 2.3 0.391 7.81 15.2 297.4 124.4 132/ 1 62 ($0.15) ($0.005) ($0.27) ($0.42) 46/78/103/ 155.8 26.6 350 2.4 0.387 7.42 14.6 317.5 101.4 135/164 ($0.16) ($0.004) ($1.07) ($0.66) ' Average values based on two measurements. Standard deviation given in parentheses. b Average value based on three measurements. Standard deviation given in parentheses. ‘ Average value based on fifteen measurements. Standard deviation given in parentheses. 74 Table A.5.4. Processing Data and Average % Thiamin Losses for CO; (Lower Temperature) and Control (Higher Temperature) Extrusion Conditions Screw % Screw C0; Melt Temp. (°C) Product "/6 Die Press. Average °/e Config. Moist. Speed Press. Temp. at Torq. (ps1) Thiamin (rpm) (99‘) Die (°C) Loss' Control 22 200 N/A 41/75/106/ 161.3 55.0 700 12.0 ($ 1.8) 138/165 Control 22 200 N/A 44/79/102/ 154.6 47.5 620 9.9 ($ 1.2) 136/164 Control 22 250 N/A 40/76/105/ 160.2 56.0 720 15.3 ($ 4.4) l40/164 Control 22 250 N/A 41/77/104/ 158.7 55.7 720 9.8 ($ 0.5) 138/164 Control 22 300 N/A 41/76/127/ 161.3 55.0 700 15.9 ($ 0.3) 143/165 Control 22 300 N/A 41/77/104/ 159.6 53.0 680 11.7 ($ 0.7) 142/163 CO; 22 300 0 41/61/71/89/96 107.1 32.5 600 <0 ($ 2.2) C02 22 300 0 41/62/71/91/97 114.7 35.9 820 1.5 ($ 0.8) CO; 22 300 150 41/64/72/85/97 1 12.7 44.0 940 <0 ($ 2.9) C02 22 300 150 41/59/69/86/93 104 32.3 600 <0 ($ 0.8) C02 22 350 41/60/71/92/97 109.2 34.5 690 <0 ($ 4.1) CO; 22 350 0 41/60/70/91/96 112.2 35.6 790 <0 ($ 1.7) C02 22 350 150 41/63/72/86/97 1 15.6 34.0 700 8.0 ($ 0.5) C02 22 350 150 41/60/69/85/93 106.6 31.8 620 11.0 ($ 0.8) C02 22 400 0 41/56/65/82/89 106.6 30.0 610 3.2 (_$ 0.3) CO; 22 400 0 41/58/67/84/91 108.2 31.2 760 6.5 ($ 0.7) CO; 22 400 150 41/65/75/88/98 1 18.8 36.0 700 1.7 ($ 2.9) CO; 22 400 150 41 /59/69/87/95 l 1 1.5 30.9 600 5.4 ($ 0.01) CO; 25 300 0 42/63/76/84/85 96.3 32.5 335 26.4 ($ 1.1) CO; 25 300 0 41 /61/71/83/89 95.9 31.5 400 0.5 ($ 0.4) CO; 25 300 150 41/63/74/87/91 99.5 34.4 530 16.1 ($ 0.3) CO; 25 300 150 41/69/85/94.’91 100.8 36.2 335 1.1 ($ 1.5) CO; 25 350 0 41/62’75/82/84 93.1 32.0 340 11.6 ($ 1.0) CO; 25 350 0 41/61/71/86/92 101.2 34.5 540 <0 ($ 1.5) C02 25 350 150 41/61/74/87/92 101.0 33.1 560 24.5 ($ 1.3) C02 25 350 150 41/70/86/95/92 103.1 34.5 330 33.8 ($ 1.1) CO; 25 400 0 41/57/68/77/85 95.2 29.3 290 28.5 ($ 0.7) CO: 25 400 0 44/59/69/84/89 96.8 34.1 560 28.1 ($ 0.6) CO; 25 400 150 44/59/69/87/92 102.5 34.0 600 28.5 ($ 1.4) C02 25 400 150 41/60/84/89/91 101.3 32.5 300 29.3 ($ 1.5) ' Average based on three measurements. Standard deviation given in parentheses. 75 Table A.5.5. Processing Data and Average % Thiamin Losses for CO; Extrusion at Higher Temperatures Screw % Screw CO; Melt Temp. Product Temp. °/o Dle Press. Average °/e Config. Moist. Speed Press. (°C) at Die (°C) Torque (psi) Thiamin (rpm) (psi) Loss' C02 22 300 0 44/75/102/ 153.4 24.9 300 2.1 $ (1.2) 132/162 C02 22 300 0 46/78/103/ 155.8 26.6 350 3.0 $ (3.4) 135/164 ' Average based on three measurements. Standard deviation given in parentheses. 76 Table A.5.6. Averages and % CVs for % Thiamin in Feed Flour and Extruded Products on a Dry Weight Basis Screw % Screw CO, Average % Average % Configuration Moisture Speed Pressure Thiamin Thiamin (rpm) (ps1) Feed Flour’ Extruded (Dry Wt.) Products' (Dry Wt.) Control 22 200 N/A 0.317 (3.7%) 0.279 (2.0%) Control 22 200 N/A 0.331 (0.7%) 0.298 (1.1%) Control 22 250 N/A 0.317 (3.7%) 0.268 (5.2%) Control 22 250 N/A 0.331 (0.7%) 0.299 (1.6%) Control 22 300 N/A 0.317 (3.7%) 0.266 (0.4%) Control 22 300 N/A 0.331 (0.7%) 0.292 (0.8%) C02 22 300 0 0.3 I 6 (6.0%) 0.364 (1.9%) C02 22 300 O 0.349 (2.6%) 0.344 (0.8%) CO; 22 350 0 0.316 (6.0%) 0.360 (2.5%) CO; 22 350 0 0.349 (2.6%) 0.354 (0.8%) CO; 22 400 0 0.316 (6.0%) 0.368 (3.5%) CO; 22 400 0 0.349 (2.6%) 0.357 (1.7%) CO; 22 300 150 0.323 (5.0%) 0.297 (0.6%) CO; 22 300 150 0.316 (6.0%) 0.281 (0.9%) CO; 22 350 150 0.323 (5.0%) 0.312 (0.3%) CO; 22 350 150 0.316 (6.0%) 0.295 (0.8%) CO; 22 400 150 0.323 (5.0%) 0.317 (2.9%) CO; 22 400 150 0.351 (1.8%) 0.333 (0.0%) CO; 25 300 0 0.320 (1.8%) 0.235 (1.4%) CO; 25 300 0 0.349 (2.6%) 0.347 (0.4%) CO; 25 350 0 0.320 (1.8%) 0.268 (0.4%) C02 25 350 0 0.349 (2.6%) 0.346 (1 .5%) C02 25 400 0 0.320 (1.8%) 0.283 (1.1%) C02 25 400 0 0.349 (2.6%) 0.351 (1.5%) CO; 25 300 150 0.327 (2.0%) 0.247 (1.7%) CO; 25 300 150 0.320 (1.8%) 0.271 (1.7%) CO; 25 350 150 0.327 (2.0%) 0.234 (3.7%) CO; 25 350 150 0.320 (1.8%) 0.261 (0.8%) CO; 25 400 150 0.327 (2.0%) 0.234 (1.7%) CO; 25 400 150 0.320 (1.8%) 0.267 (2.2%) ' Average based on three measurements. % CV given in parentheses. 77 Table A.5.7. Averages and % CVs for % Thiamin in Feed Flour and Extruded Products on a Dry Weight Basis, Using CO; Extrusion at Higher Temperatures Screw °/o Screw CO, Average °/o Average % Configuration Moisture Speed Pressure Thiamin Thiamin (rpm) (psi) Feed Flour“ Extruded (Dry Wt.) Products' 1D!) Wt.) CO; 22 300 0 0.349 (2.6%) 0.342 (1.2%) CO; 22 300 0 0.349 (2.6%) 0.334 (3.5%) ' Average based on three measurements. % CV given in parentheses. 78 Table A.5.8. Example: Mean Residence Time Calculation, High Temperature (Control) at a Screw Speed of 200 rpm Norm. Rel. Conc. Conc. t ]|C(t)| delta t} |C(t)| deltat Time (5) 9(1): E_(_t_) t delta t (Mid_dle Time Weigm (Total Dye Amt.) 0 0 0.0000 0 0 0 0 41.97 10.8 0.0236 44.47 5 2401.38 54 46.97 11.6 0.0253 49.47 5 2869.26 58 51.97 10.6 0.0231 54.47 5 2886.91 53 56.97 10.2 0.0223 59.47 5 3032.97 51 61.97 8.9 0.0194 64.47 5 2868.915 44.5 66.97 7.5 0.0164 69.47 5 2605.125 37.5 71.97 6.4 0.0140 74.47 5 2383.04 32 76.97 4.9 0.0107 79.47 5 1947.015 24.5 81.97 3.9 0.0085 84.47 5 1647.165 19.5 86.97 3 .2 0.0070 89.47 5 1431.52 16 91.97 2.6 0.0057 94.47 5 1228.11 13 96.97 2 0.0044 99.47 5 994.7 10 101.97 1.8 0.0039 104.47 5 940.23 9 106.97 1.1 0.0024 109.47 5 602.085 5.5 111.97 1.1 0.0024 114.47 5 629.585 5.5 116.97 1.1 0.0024 119.47 5 657.085 5.5 121.97 1 0.0022 124.47 5 622.35 5 126.97 1 0.0022 129.47 5 647.35 5 131.97 1.1 0.0024 134.47 5 739.585 5.5 136.97 0.8 0.0017 139.47 5 557.88 4 141.97 0.4 0.0009 56.788 0 Total 31749.048 458 Mean RTD (Seconds) 69.32 ' C(t) adjusted to zero based on control sample with no dye. 79 Table A.5.9. Results of MS Excel t-Test Comparing Mean Expansion Ratios at 0 and 150 psi C02, for a Given Screw Speed and at 22% Moisture t-Test: Paired Two Sample for Means at 300 rpm (no sigpificant difference at p <0.05) 15031 C02 0 psi C02 Mean 1.653 1. 718 Variance 0.126 0.178 Observations 30.000 30.000 Pearson Correlation -0.015 Hypothesized Mean Difference 0.000 df 29.000 t Stat -0. 635 P(T<=t) one-tail 0.265 t Critical one-tail 1.699 P(T<=t) two-tail 0.530 t Critical two-tail 2.045 t-Test: Paired Two Sample for Means at 350 rpm (significantly different at p <0.05) 150 psi C02 0 psi C02 Mean 1.897 2.287 Variance 0.178 0.187 Observations 30.000 30.000 Pearson Correlation -0.413 Hypothesized Mean Difference 0.000 df 29.000 t Stat -2. 971 P(T<=t) one-tail 0.003 t Critical one-tail 1.699 P(T<=t) two-tail 0.006 t Critical two-tail 2.045 t-Test: Paired Two Sample for Means at 400 rpm (no significant difference at p <0.05) ISMSI C02 0 psi C02 Mean 2.213 2.277 Variance 0.061 0.026 Observations 30.000 30.000 Pearson Correlation -0.035 Hypothesized Mean Difference 0.000 df 29.000 t Stat -1.1 5 6 P(T<=t) one-tail 0. 128 t Critical one-tail 1.699 P(T<=t) two-tail 0.257 t Critical two-tail 2.045 80 Table A.5.10. Results of MS Excel t-Test Comparing Mean % Water Solubility in the High- and Low-Temperature Samples at 22% Moisture, 0 psi C02, and Screw Speed of 300 rpm t-Test: Paired Two Sample for Means at 300 rpm (no significant difference at p <0.05) High-TemperaturgControl) Low-Temperature Mean 13.565 12 Variance 0.336 2.82 Observations 4.000 4 Pearson Correlation 0.695 Hypothesized Mean Difi‘erence 0.000 df 3.000 t Stat 2.272 P(T<=t) one-tail 0.054 t Critical one-tail 2.353 P(T<=t) two-tail 0.108 t Critical two-tail 3.182 Table A.5.11. Results of MS Excel t-Test Comparing Mean % Water Solubility in Low-Temperature Samples at 22% and 25% Moisture t-Test: Paired Two Sample for Means at 300 rpm, 0 psi C02 at Lower Temperatures (significantly different at p <0.05) 22% Moisture 25% Moisture Mean 12.000 6.675 Variance 2.820 0.289 Observations 4.000 4.000 Pearson Correlation -0.7 38 Hypothesized Mean Difference 0.000 df 3.000 t Stat 5. 053 P(T<=t) one-tail 0.007 t Critical one-tail 2.3 53 P(T<=t) two-tail 0.015 t Critical two-tail 3.182 81 REFERENCES 82 AACC. 2000. Association of Cereal Chemists Approved Methods, 10‘h ed., St. Paul, MN Altomare, RE, and Ghossi, P. 1986. An analysis of residence time distribution patterns in a twin screw cooking extruder. Biotech. Progress. 2(3): 157-163 Anderson, R.A., Conway, H.F., Pfeifer, V.F., and Griffin, E.L., Jr. 1969. Gelatinization of corn grits by roll- and extrusion-cooking. Cereal Sci. Today. 14: 4-12 AOAC International. 1995. Official Methods of Analysis, 16th ed., AOAC International, Arlington, VA. Asp, N.G., and Bj6rck, I. 1989. Nutritional properties of extruded foods. In Extrusion Cooking, eds. C. Mercier, P. Linko, and J. M. Harper, pp. 399-434. AACC, St. Paul, MN Association of Vitamin Chemists. 1966. Methods of Vitamin Assay, 3rd ed., pp.123-145. Interscience Publishers, New York Beetner, G., Tsao, T., Frey, A., and Harper, J. 1974. Degradation of thiamine and riboflavin during extrusion processing. J. Food Science. 39: 207-208 Bhattacharya, M., and Hanna, M.A. 1987. Textural properties of extrusion-cooked corn starch. Lebensm. Wiss. T echnol., 20 (4): 195-201 Bhattacharya, S., and Choudury, GS. 1994. Twin-screw extrusion of rice flour: effect of extruder length-to-diameter ratio and barrel temperature on extrusion parameters and product characteristics. J. Food. Proc. Pres. 18(5): 389-406 Bj6rck, I., and Asp, N.G. 1983. The effects of extrusion cooking on nutritional value—a literature review. J. Food Engineering. 2: 281-308 Brubacher, G., Milller-Mulot, and Southgate, D.A.T. 1985. In Methods for the Determination of Vitamins in Food. Elsevier Applied Science Publishers. London, p. 51- 65 Burns, R.E., Caldwell, E.F., Fast, RB, and Jones, W.H. 2000. Application of nutritional and flavoring/sweetening coatings. In Breakfast Cereals and How They Are Made, 2” ed., eds. R.B. Fast and E. F. Caldwell, pp. 279-313, AACC, St. Paul, MN Chiang, B. -Y., and Johnson, J .A. 1977. Gelatinization of starch in extruded products. Cereal Chem. 54(3): 436-443 Colonna, P., Doublier, J.L., Melcion, J .P., de Monredon, F., and C. Mercier. 1984. Extrusion cooking and drum drying of wheat starch. 1. Physical and macromolecular modifications. Cereal Chem. 61(6): 538-543 83 Colonna, P., Tayeb, J ., and Mercier, C. 1989. Extrusion cooking of starch and starchy products. In Extrusion Cooking, ed. C. Mercier, P. Linko, and J. M. Harper, pp. 247-320. AACC, St. Paul, MN Dwivedi, B.K., and Arnold, R.G. 1973. Chemistry of thiamine degradation in food products and model food systems: A review. J. Agr. Food Chem. 21(1): 54-60 Ferdinand, J .M., Lai-Fook, R. A., Ollett, A. L., Clark, S. A., Smith, AC. 1990. Structure formation by carbon dioxide injection in extrusion cooking. J. Food Engineering. 11: 209-224 Ferdinand, J. M., Clark, S. A., Smith, A. C. 1992. Structure formation in extrusion- cooked starch-sucrose mixtures by carbon dioxide injection. J. Food Engineering. 16: 283-291 Food and Drug Administration. 1994. Guide to Nutrition Labeling and Education Act (NLEA) Requirements. Gubler, C.J. 1991. Thiamin. In Handbook of Vitamins, 2nd ed., ed. L.J. Machlin, pp. 233-281. Marcel Dekker, Inc., New York Guzrnan-Tello, R., and Chefiel, J .C. 198 7. Thiamine destruction during extrusion cooking as an indicator of the intensity of thermal processing. International J. Food Science and Tech. 22: 549-562 Hankansson, B., Jagerstad, M., Oste, R., Akesson, B., and Jonsson, L. 1987. The effects of various thermal processes on protein quality, vitamins, and selenium content in whole- grain wheat, and white flour. J. Cereal Sci. 6: 269-282 Harper, J .M. 1978. Extrusion processing of food. Food Technology. 32(7): 67-72 Harper, J. M. 1981. Extrusion of foods. Vol. 1. CRC Press. Boca Raton, Florida Harper, J.M. 1988. Extrusion processing. In Nutritional Evaluation of Food Processing, 3rd ed., eds. E. Kaunas and RS. Harris, pp. 365-391. Van Nostrand Reinhold Company, New York Hauck, B.W., and Huber, GR. 1989. Single screw vs. twin screw extrusion. Cereal Foods World. 34(11): 930-939 Hernnann, F. and Tunger, L. 1966. Thermal destruction of thiamin in relation to moisture content of foods, with special reference to flour products. Nahrung. 10 (8): 705-712 (German) 110, S., and Berghorfer, E. 1998. Kinetics of themromechanical destruction of thiamin during extrusion cooking. J. Food Science. 63(2): 312-316 84 Kamman, J.F., Labuza, TR, and Warthesen, J .J . 1980. Thiamin and riboflavin analysis by high performance liquid chromatography. J. Food Science. 45(6): 1497-1499, 1504 Killeit, U. 1994. Vitamin retention in extrusion cooking. Food Chemistry. 49: 149-155 Kokini, J .L., Chang, C. N., and Lai, L. S. 1992. The role of rheological properties on extrudate expansion. In Food Extrusion Science and Technology, ed. J .L. Kokini, C.T. Ho, and M.V. Karwe, pp.631-652. Marcel Dekker, Inc., New York Labuza, TR, and Riboh, D. Theory and application of Arrhenius kinetics to the prediction of nutrient losses in foods. Food Technology. 36(10): 68-70, 72, 74 Lee, E.Y., Ryu, G. H., Lirn S. T. 1999. Effects of processing parameters on physical properties of cornstarch extrudates expanded using supercritical carbon dioxide injection. Cereal Chem. 76(1):63-69 Levenspiel, O. 1999. Chemical Reaction Kinetics. 3rd ed. John Wiley & Sons, Inc., New York. Maga, J .A., and Sizer, CE. 1978. Ascorbic acid and thiamin retention during extrusion of potato flakes. Lebensm. Wiss. Tech. 11: 192-194 Mason, W.R., and Hoseney, R. C. 1986. Factors affecting the viscosity of extrusion- cooked wheat starch. Cereal Chem. 63(5): 436-441 Mercier, C., and Feillet, P. 1975. Modification of carbohydrate components by extrusion cooking of cereal products. Cereal Chem. 52(3): 283 Mohamed, I.O., Ofoli, R.Y., and Morgan, R.G. 1990. Modeling the average shear rate in a co-rotating twin screw extruder. J. Food Proc. Engineering. 12: 227-246 Mulley, E.A., Stumbo, CR, and Hunting, W.M. 19758. Kinetics of thiamine degradation by heat. Effect of pH and form of the vitamin on its rate of destruction. J. Food Science. 40: 989-992 Mulley, E.A., Stumbo, CR, and Hunting, W.M. 1975b. Kinetics of thiamine degradation by heat. A new method for studying reaction rates in model systems and food products at high temperatures. J. Food Science. 40: 985-988 Mulvaney, S. J ., and Rizvi, S.S.H. 1993. Extrusion processing with supercritical fluids. Food Technology. 47: 74-82 85 Pham, C. B., and Del Rosario, RR. 1986. Studies on the development of texturized vegetable products by the extrusion process. 111. Effects of processing variables on thiamin retention. J. Food Tech. 21: 569-576 Rindi, G. 1996. Thiamin. In Present Knowledge in Nutrition, 7th ed., eds. Ziegler, and E. Ekhard, L.J. Flier, Jr., pp. 160-166. ILSI Press, Washington, DC Rizvi, S.S.H., Mulvaney, S.J., and Sokhey, AS. 1995. Trends in Food Science and Technology. 6: 233-240 Sloan, A.E. 2000. The top ten firnctional food trends. Food Technology. 54(4): 33-62 Sloan, A.E. 2001. Move over, chips and dip. Food Technology. 55(9): 20 Smith, IS 1994. Evaluation of analytical data. In Introduction to the Chemical Analysis of Foods, ed. S. S. Nielsen, pp. 53-55, Jones and Bartlett Publishers. London/Boston Sokhey, A.S., Rizvi, SSH, and Mulvaney, SJ. 1996. Application of supercritical fluid extrusion to cereal processing. Cereal Foods World. 41(1): 29-34 Toma, R. B. and Tabekhia, M.M. 1979. High performance liquid chromatographic analysis of B-vitamins in rice and rice products. J. Food Sci. 44(1): 263-266, 268 Vainionpiiii, J. 1991. Modelling of extrusion cooking of cereals using response surface methodology. J. Food Engineering. 13: 1-26 86 1111111111111111