14 extrusion cooking

Table 14.1 Common food products prepared by extrusion cookingReady-to-eat breakfast cereals Puffed cereals Flaked cereals High-fiber strands Half-products or pellets third generation sna

Extrusion cooking

M E Camire, University of Maine

14.1 Introduction

Extrusion cooking is a relatively recent form of food processing Forcing material through a hole is the process of extrusion Sausage extruders were devel-oped in the nineteenth century as simple forming machines Eventually pasta was produced in extruders Flour and water were added at one end of the machine, and a screw mixed and compressed the dough before extruding it through numer-ous holes or dies that gave the pasta its shape During the 1930s heat was added

to the barrel containing the screw; puffed corn curl snacks resulted The pressure developed as the dough moved along the screw; this, together with the heat under pressure, caused the corn to puff upon exiting the dies As extrusion cooking processed more types of food, extruders became more specialised for food applications Twin-screw extruders containing two screws were adapted from the polymer industry, and these machines are considerably more versatile than are single screw extruders Extruded products are often subjected to further process-ing, such as fryprocess-ing, bakprocess-ing, and rolling

The improved mixing ability of these extruders provided impetus for further product development Table 14.1 lists major food categories produced by extru-sion cooking Extruextru-sion cooking can be performed as either a batch or con-tinuous operation, offering many advantages over conventional food processing methods (Table 14.2) (Harper, 1981) Several manufacturers produce cooking extruders Laboratory-size extruders have screw diameters of 10–30 mm and throughputs of up to a few hundred kilogram per hour The length of the barrels for these research extruders, which are the most common machines cited in the literature, varies from about one to two meters Production-sized extruders can create thousands of kilogram of product per hour

Table 14.1 Common food products prepared by extrusion cooking

Ready-to-eat breakfast cereals Puffed cereals

Flaked cereals High-fiber strands

Half-products or pellets (third generation snacks) Crispbreads

Chocolate Texturised protein Soy meat analogues

Restructured seafood Processed cheese

Weaning cereals

Table 14.2 Unique advantages of extrusion cooking a

Batch or continuous processing

High throughput

Low labor and energy costs

Variety of products produced and types of ingredients that

can be processed

Control of thermal/mechanical environment

Negligible effluent

a Adapted from Harper (1981).

14.1.1 Unique aspects of extrusion cooking

Most extruders act as heat exchangers, and they also shape and form food products Mixing, dehydration, and pasteurisation and sterilisation are other unit operations that typically occur during extrusion Aside from thermal destruction

of nutrients, the shear that develops within the extruder barrel can damage food chemicals Temperature can be controlled by many means including limiting direct heating, adding water, and increasing throughput Shear may be reduced by using low-shear screw elements, increasing water or lipid content, modifying screw speed (based on other parameters), and by reducing pressure at the die Extrusion research often focuses upon one to four variables, although screen-ing studies should be performed to identify key factors Extruder operators may select parameters such as screw speed, feed moisture, and barrel temperature as primary factors, that in turn determine a secondary set of factors: specific mechan-ical energy (SME), product or mass temperature (PT) and pressure (Meuser and van Lengerich, 1984) These factors influence the viscosity of the food within the extruder barrel, the residence time of the material in the extruder, and the shear applied to the food (Fig 14.1) Variations caused by feed composition

and prior processing of the feed materials are important sources of experimental variation

Extrusion can produce safe, lightweight, shelf-stable foods that can be stored for use during famines and natural disasters Simple single screw extruders are fairly inexpensive and simple to maintain so these machines can be used in less-developed nations to produce weaning and other foods Harper and Jansen (1985) have reviewed advantages and limitations of extrusion for weaning foods Fric-tion from the rotaFric-tion of the screw can cook the food thoroughly, reducing pro-duction costs for fuel sources Extruders can blend diverse ingredients, permitting government and relief agencies to use donated foods such as dried milk as well

as indigenous crops such as beans, millet, and cassava Extruded pellets can be ground, then mixed with milk or water as needed to form gruel for infants Functional ingredients such as soy and botanicals that are relatively unpalat-able alone can be incorporated into new food items by extrusion Traditional foods such as rye crispbread can be further enhanced by addition of extra dietary fiber or other ingredients during extrusion A relatively new form of extrusion known as wet extrusion operates at higher moisture contents (>40%) and lower

Primary extrusion factors

Extruder model

Feed composition, particle size, preconditioning

Feed rate

Added water rate

Barrel temperature

Screw configuration and speed

Die number and geometry

Secondary extrusion factors

Mass or product temperature Viscosity

Pressure Specific mechanical energy

Nutrient changes

Retention Destruction Bioavailability

Fig 14.1 Interrelationships of extruder variables and their potential effects on nutrients.

barrel temperatures (Akdogan, 1999) These conditions permit extrusion and texturisation of high-protein materials since protein denaturation is limited Very little has yet been published on the effects of wet extrusion on nutrient retention, but nutrient destruction should be considerably less than in conventional extru-sion cooking

14.2 Impact on key nutrients: carbohydrates

Reducing sugars such as glucose and lactose participate in Maillard reactions, which will be discussed further in section 14.3 The shear forces during extru-sion can also create reducing sugars from complex carbohydrates as well as from sucrose and other sugars Sucrose losses of up to 20% were found in protein-enriched biscuits (Noguchi and Cheftel, 1983) While sucrose loss may affect product color and flavor, there is an opportunity to reduce the content of indigestible oligosaccharides that can cause flatulence Sucrose, raffinose and stachyose decreased significantly in extruded pinto bean high-starch fractions (Borejszo and Khan, 1992) Corn-soy snacks had lower levels of both stachyose and raffinose compared to unextruded soy grits and flour, but values were not corrected for the 50–60% corn present (Omueti and Morton, 1996) Starch and stachyose were lower in extruded peas compared to raw peas (Alonso et al, 2000), but an increase in total free sugars did not fully account for these losses (Fig 14.2)

500

450

400

350

300

250

200

150

100

50

0

Starch Stachyose Total free sugars

*

*

*

Raw Extruded

Fig 14.2 Carbohydrate changes (g/kg dry matter) due to extrusion of peas (Pisum sativum L) at an exit temperature of 145 °C and 25% feed moisture (Adapted from Alonso

et al, 2000)

Starch is usually the major food constituent in extruded foods such as break-fast cereals, snacks and weaning foods Humans and other monogastric species

do not readily digest native or ungelatinised starch Unlike many thermal processes, extrusion cooking gelatinises starch at fairly low (12–22%) moisture levels Removal of cooking water is not a problem, and leaching of water-soluble nutrients is avoided Increased temperature, shear, and pressure during extrusion increase the rate of gelatinisation, but lipids, sucrose, dietary fiber and salts can retard gelatinisation (Jin et al, 1994) While full gelatinisation may not occur during extrusion, digestibility is often improved (Wang, S et al, 1993)

During extrusion, starch molecules can be physically broken into smaller, more digestible fragments For example, amylopectin branches can be sheared off the main molecule, with larger molecules experiencing the greatest effect (Politz et al, 1994b) Both amylose and amylopectin molecules may be affected, however Molecular weight in extruded wheat starch was retained better under processing conditions of higher die temperature (185 °C) and feed moisture (20%) (Politz et al, 1994a) Screw configurations using more reverse and high-shear ele-ments favor starch breakdown (Gautam and Choudhoury, 1999)

Lower molecular weight starch fragments may be sticky, thereby increasing the risk for dental caries, since bacteria in the mouth rapidly ferment these dextrins Toothpack, the amount of material retained on teeth, has been used as

an indication of the severity of extrusion processing Björck and co-workers (1984) found that white wheat flour extruded under ‘mild’ and ‘severe’ condi-tions caused drops in dental plaque pH comparable to those obtained with glucose

While easily-digested starch is desirable for infants and invalids, the resulting rapid post-prandial rise in blood sugar and insulin levels is thought to be a risk factor for development of insulin insensitivity and Type II, or adult-onset, diabetes Extrusion offers the ability to reduce the high glycemic index (GI) of some foods by converting starch to digestion-resistant starch (RS) Theander and Westerlund (1987) reported transglycosidation in extruded wheat flour, presum-ably from attachment of sheared amylopectin branches to other reactive sites The resulting novel bonds would be resistant to digestion by enzymes Addition of high amylose starch also reduces digestibility As much as 30% resistant starch was reported when high amylose starch was reacted with pullulanase prior to extrusion (Chiu et al, 1994) Extruded high amylose rice noodles had lower starch digestibility and reduced GI (Panlasigui et al, 1992)

An evolving area of research involves the use of additives to promote RS for-mation Adding 30% corn, potato or wheat starch did not increase RS values in cornmeal, but RS and fiber values more than doubled when 7.5% citric acid was used, and 30% high-amylose cornstarch with 5 or 7.5% citric acid resulted in values of 14%, compared with slightly more than 2% in 100% cornmeal (Unlu and Faller, 1998) Polydextrose may have been formed during extrusion Limi-tations to this approach would be the expenses of the additives and sour taste of the extrudates Yields of up to 93.7% oligosaccharides and polydextrose were

reported when glucose-citric acid mixtures were extruded at different barrel tem-peratures (Hwang et al, 1998)

Longer cellulose fibers added to cornstarch decreased starch solubility (Chinnaswamy and Hanna, 1991) Removal of insoluble dietary fiber from wheat flour in combination with 20% protein addition resulted in pasta with

significantly delayed dextrin release under in vitro digestion conditions (Fardet

et al, 1999), possibly due to enhanced protein–starch interactions Amylose forms complexes with lipids during extrusion, thereby reducing both starch and lipid availability This phenomenon will be addressed in section 14.4

The term dietary fiber is used to describe nondigestible carbohydrates and associated compounds such as lignin Although a global definition of dietary fiber does not yet exist, there is a consensus that adequate fiber consumption is essen-tial for good health Analytical methods for quantitating dietary fiber vary con-siderably The AOAC total dietary fiber method used for US nutritional labeling does not measure compounds that are soluble in 80% aqueous ethanol such as certain fructans and polydextrose, and this procedure does not detect changes in extruded fiber solubility If different dietary fiber fractions are not analysed sepa-rately, it is possible to overlook important changes in dietary fiber composition and functionality caused by extrusion

Like starch, branched dietary fiber molecules are susceptible to shear The smaller fragments may be soluble in water Fragments may also combine to form large insoluble complexes that may be analysed as lignin Although extrusion did not affect pectin, both soluble and insoluble nonstarch polysaccharides (NSP) were increased in extruded oatmeal and potato peels (Camire and Flint, 1991) Corn meal fiber was unaffected by extrusion under the same conditions as the other

foods Extruded beans (Phaseolus vulgaris L) had total fiber values comparable to

those before extrusion, but a redistribution of insoluble to soluble fiber occurred (Martín-Cabrejas et al, 1999) Sugar beet pectin and hemicellulose molecular weight decreased with extrusion, and water solubility of those compounds in-creased by 16.6 to 47.5% (Ralet et al, 1991) Extrusion inin-creased the solubility of beta-glucans in regular and waxy barley cultivars (Gaosong and Vasanthan, 2000) Does the ‘soluble’ fiber created during extrusion have the same health ben-efits as natural forms such as pectin and b-glucan? Viscous gels formed in the small intestine trap bile acids and thus may contribute to lower serum cholesterol levels; the soluble fiber matrix is also thought to slow glucose absorption from the small intestine Extrusion increased the viscosity of aqueous suspensions of

wheat, oats and barley (Wang and Klopfenstein, 1993) Although increased in

vitro viscosity was correlated with higher levels of soluble citrus peel fiber after

extrusion (Gourgue et al, 1994), in vitro starch digestion and glucose diffusion

were unaffected Extrusion of wheat flakes containing guar gum did not reduce the guar gum’s ability to lower post-prandial blood glucose and insulin in healthy adults (Fairchild et al, 1996) In an intervention study involving middle-aged men with hyperlipidemia, baked goods fortified with 92 g/day extruded dry white beans did not lower serum lipoproteins (Oosthuizen et al, 2000)

14.3 Proteins

Two reviews of protein extrusion have been published (Camire, 1991; Arêas, 1992) The effects of extrusion on protein nutrition have been studied extensively for animal feeds and for human weaning foods Total protein changes very little during most extrusion operations Changes in nutritional quality may be

over-looked if only total nitrogen is assayed; animal feeding studies or in vitro protein

digestibility testing should be performed on products that are designed to provide significant amounts of high-quality protein Disulfide and other covalent cross-linking, aggregation and fragmentation are among reactions reported in the litera-ture Free radical reactions are significant during wheat flour extrusion (Schaich and Rebello, 1999)

Excessive Maillard browning can result in losses of lysine up to approximately 50% (de la Gueriviere et al, 1985) High barrel temperature, low moisture, and high shear promote Maillard reactions Browning may occur even when reduc-ing sugars are excluded from formulations because new reducreduc-ing sugars may be formed from hydrolysis of sucrose, starch, and other polysaccharides In a model system of wheat starch, glucose and lysine, low pH increased Maillard reactions (Bates et al, 1994) Lysine can be preserved, however, if extruder operating conditions and formulations are carefully balanced Corn-soy blends extruded for reconstitution as porridge or gruel had good lysine retention (Konstance et al, 1998)

Extrusion may improve protein digestibility by denaturating proteins, ex-posing enzyme-accessible sites Enzymes and enzyme inhibitors generally lose activity due to denaturation Reductions in protease inhibitors can contribute to better plant protein utilisation Although a single test for protein denaturation is not used internationally, protein solubility in water or aqueous solutions is com-monly used to assess the extent of denaturation High shear extrusion conditions

in particular promote denaturation (Della Valle et al, 1994), although mass tem-perature and moisture are also important factors Protein solubility is reduced in pasta despite the low process temperatures used in pasta making (Ummadi et al, 1995)

The mechanism for cholesterol lowering by a diet with soy protein is not well understood, but the lysine/arginine ratio may play an important role Health effects of food proteins could be significantly affected by extrusion cooking if lysine is selectively lost via Maillard reactions Extrusion-texturised soy isolate fed to rats had similar effects as nonextruded soy on serum cholesterol, choles-terol and steroid fecal excretion, or protein nutrition (Fukui et al, 1993) In another

rat study, amino acid-supplemented extruded pea (Pisum sativum L., cv Ballet)

seed meal lowered total and LDL cholesterol as well as did supplemented raw seeds compared with a control diet (Alonso et al, 2001) The peas were extruded under fairly mild conditions (145 °C exit temperature and feed moisture of 25%), but antinutritional factors were adequately inactivated, as evidenced by lower pancreatic weights in rats fed the extruded peas Amaranth protein has a lysine/arginine ratio similar to soy LDL cholesterol in rabbits fed an extruded

amaranth diet for 21 days was less than half that in animals fed a casein control diet (Plate and Areâs, 2002)

14.4 Lipids

Extruded foods are generally low in lipid content, but fat is often added post-extrusion by frying or spraying of lipids to hold seasonings Generally, foods containing less than 10% lipids are extruded because greater quantities of lipids reduce slip within the extruder barrel, making extrusion difficult, particularly for expanded products Single screw extruders can process lipid levels of 12–17%, while twin screw extruders with proper screw configurations can handle feed lipid contents as high as 22% (Riaz, 2001) Extruders are used in oilseed extraction because the heat and shear disrupt cellular tissue and free oil (Nelson

et al, 1987)

An early problem in extrusion was the apparent disappearance of lipids during processing Starch–lipid complexes formed during extrusion are resistant to some lipid extraction procedures Lipid recovery is higher when extruded foods are first digested with acid or amylase, then extracted with ether or another organic solvent While total fat was not significantly changed in extruded whole wheat, just half of ether-extractable lipids were detected (Wang, W-M et al, 1993) In the same study, wheat bran had more free lipids after extrusion Cornmeal extruded at lower barrel temperatures (50–60 °C or 85–90 °C) had greater than 75% of its lipids bound, but extrusion at 120–125 °C bound 70% of the lipids (Guzman et al, 1992)

Other nutritional aspects of lipids before and after extrusion have been studied very little Both docosahexaenoic (DHA) and eicosapentaenoic (EPA) acids were retained in extruded chum salmon muscle with 10% wheat flour (Suzuki et al, 1988) Unlike other processing methods, extrusion cooking does not promote

significant cis-trans isomerisation of unsaturated lipids Corn and soy blends had 1.5% more trans-fatty acids after extrusion (Maga, 1978) Formation of free

radicals and subsequent lipid oxidation could have nutritional implications Artz

et al (1992) reviewed extrusion factors affecting lipid oxidation Screw and barrel wear raise levels of pro-oxidant minerals in extruded foods For example, iron and peroxide values were higher in extruded rice and dhal compared to dried products (Semwal et al, 1994) Increased surface area in expanded products is another factor increasing oxidation Factors that retard oxidation in extruded foods include denaturation of lipolytic enzymes, formation of starch-lipid com-plexes, and creation of antioxidant Maillard compounds

14.5 Vitamins

Killeit (1994) reviewed vitamin retention in extruded foods More research on the bioavailability of added and endogenous vitamins is needed, particularly in

light of fortification programs for folate and other vitamins Concerns of reduced vitamin levels prompt some manufacturers to apply vitamins post-extrusion as a spray More recent research has focused on vitamin stability in feeds Fat-coated ascorbic acid, menadione, pyridoxine and folic acid were retained better than were crystalline forms in extruded fish feed (Marchetti et al, 1999)

Although many extruded foods do not naturally have high levels of lipid-soluble vitamins, stability of these nutrients is a concern for fortified foods Over

50% of all trans-beta-carotene in wheat flour were destroyed when barrel

tem-perature increased from 125 to 200 °C (Guzman-Tello and Cheftel, 1990) The degradation process is not straightforward Fifteen degradation products of all

trans-beta-carotene dispersed in corn starch were recovered after twin-screw

extrusion (Marty and Berset, 1988) Retention of retinyl palmitate in tapioca snacks mixed with either fish or protein flour was 52% and 73%, respectively after extrusion (Suknark et al, 2001)

Vitamins D and K are fairly stable during food processing, but are not used

in many extruded human foods Vitamin E and related tocopherols function as both vitamin and antioxidant Gamma and delta tocopherols underwent greater losses (~40%) during extrusion than did alpha and beta forms (23–28%) (Suknark

et al, 2001) Rice bran tocopherol decreased as extrusion temperature increased; bran extruded at 120–140 °C lost more tocopherols over a year’s storage than did bran extruded at 110 °C (Shin et al, 1997) Less than 20% of vitamin E was retained in extruded and drum-dried wheat flour (Wennermark, 1993) The sta-bility of lipid-soluble vitamins is shown in Table 14.3

Ascorbic acid (vitamin C) decreased in wheat flour when extruded at higher barrel temperatures at fairly low moisture (10%) (Andersson and Hedlund, 1991) Blueberry concentrate appeared to protect 1% added vitamin C in an extruded breakfast cereal compared to a product containing just corn, sucrose, and ascor-bic acid (Chaovanalikit, 1999) When ascorascor-bic acid was added to cassava starch

Table 14.3 Stability of lipid-soluble vitamins during the

production of tapioca–peanut flour snacks

Processing step Vitamin content (g/100 g), fat- and

moisture-free basis Total tocopherols Retinyl palmitate Raw material 14.44 b

2.74 a

Extrusion 10.55 c

2.19 b

Drying to form 10.36 c

1.92 c

half-product

Frying 43.18 a

2.00 c

Different letters within columns indicate statistically significant

dif-ferences.

Adapted from Suknark et al, 2001.

to increase starch conversion, retention of over 50% occurred at levels of 0.4–1.0% addition (Sriburi and Hill, 2000)

Thiamine is the water-soluble vitamin most susceptible to thermal processing Thiamine destruction in extruded wheat flour is a first-order reaction (Guzman-Tello and Cheftel, 1987) Killeit (1994) summarised thiamine losses as ranging from 5 to 100% Thiamine retention in potato flakes decreased under extrusion conditions of lower moisture and higher barrel temperature; sulfites in the potato flakes may have also contributed to vitamin destruction (Maga and Sizer, 1978) Large losses of thiamine occurred when no water was added during extrusion, but riboflavin (B2) and niacin were not affected (Andersson and Hedlund, 1991) Using low-cost single screw extruders, Lorenz and Jansen (1980) found reten-tion of over 90% for thiamin, riboflavin, vitamin B6and folic acid in corn-soy blends processed at 171 °C

14.6 Minerals

Mineral content and bioavailability are generally retained well during extrusion Abrasive foods, such as brans rich in dietary fiber, or with low lipid and mois-ture content, gradually wear away metal from the extruder screws and barrel The equipment must be replaced or refurbished periodically due to this wear, as the metal accumulates in the extruded food As barrel temperature increased during single screw extrusion of potato flakes, iron content also increased (Maga and Sizer, 1978) Total iron increased by as much as 38% due to extrusion (Camire

et al, 1993) On the other hand, cornmeal, which has a low dietary fiber content, had no changes in total, elemental, or soluble iron after twin screw extrusion (Camire and Dougherty, 1998)

Although iron from screw wear is typically in the elemental form, the bioavail-ability appears adequate as long as excessive amounts of iron and related metals are not present Rats fed extruded corn and potato absorbed iron well (Fairweather-Tait et al, 1987) Utilisation of iron and zinc from wheat bran and wheat in adult human volunteers was not affected by extrusion (Fairweather-Tait

et al, 1989) Extrusion slightly increased iron availability in corn snacks under in

vitro digestion followed by dialysis (Hazell and Johnson, 1989) Low-shear

extru-sion retained dialysable iron in navy beans, lentils, chickpeas and cowpeas better than did high-shear extrusion (Ummadi et al, 1995) Weaning food blends of pearl millet, cowpea and peanut had greater iron availability and protein digestibility compared to similar foods processed by roasting (Cisse et al, 1998) None of the processed blends provided adequate iron to meet infant needs, however Zinc bioavailability of semolina and soy protein concentrate blends (85:15) (Kang, 1996) was unaffected by extrusion

Mineral bioavailability may be improved in extruded foods if mineral-binding phytate is reduced during processing Published research has had mixed results Extrusion reduced phytate levels in wheat flour (Fairweather-Tait et al, 1989), possibly due to inactivation of phytases during extrusion Although phytic acid