Journal of Cereal Science 41 (2005) 221237 www.elsevier.

com/locate/jnlabr/yjcrs

Review

Fractionation of wheat and wheat our into starch and gluten: overview of the main processes and the factors involved
Anne Van Der Borght*, Hans Goesaert, Wim S. Veraverbeke, Jan A. Delcour
Laboratory of Food Chemistry, Katholieke Universiteit Leuven, Kasteelpark Arenberg 20, B-3001 Leuven, Belgium Received 23 June 2004; revised 27 September 2004; accepted 28 September 2004

Abstract The starch and gluten components of wheat our or whole wheat kernels can be separated by a number of industrial processes. This review provides a summary of these processes from both starting materials. The wheat constituents of importance in the fractionation processes are briey introduced, and the different fractionation processes described with emphasis on the parameters affecting the separation, such as our composition, mixing and washing water, processing aids (with an emphasis on enzymes) and kernel pre-treatment (pearling) in the case of our fractionation and steeping conditions and processing aids in the case of whole wheat. Although fractionation of our is the basis for the current industrial processes, starch yields are impaired by starch damage as a result of milling and loss of starch to milling streams. On the other hand fractionation of whole kernels often leads to impaired gluten production as a result of harsh process conditions which devitalise the gluten. q 2004 Elsevier Ltd. All rights reserved.
Keywords: Wheat starch/gluten separation; Review; Processing aids; Processing conditions; Wheat constituents

1. Introduction The global wheat production in 2001 approached 600 million tons (FAO, 2003). A minor, but industrially very important and growing use of wheat is as a source of gluten and starch. This is related to the widespread industrial and food applications of wheat gluten and starch and its derivatives. World wheat starch production was ca. 4.1 million tons in 2000 (LMC International Ltd, 2002), originating from ca. 8 million tons of wheat. In the context of this review, gluten is the fraction isolated from wheat, which is enriched in gluten proteins (i.e. gliadins and glutenins). Laboratory and industrially prepared wheat gluten contains both protein and non-protein material including lipids (ca. 3.56.8%), minerals (ca.
Abbreviations: AGP, arabinogalactan-peptides; AX, arabinoxylan(s); MW, molecular weight(s); NSP, non-starch polysaccharides; S-AX, solubilised arabinoxylan(s); WE-AX, water extractable arabinoxylan(s); WEF, water extractable fraction; WU-AX, water unextractable arabinoxylan(s); XAA, xylanase from Aspergillus aculeatus; XBS, xylanase from Bacillus subtilis. * Corresponding author. Tel.: C32 1632 1634; fax: C32 1632 1997. E-mail address: anne.vanderborght@agr.kuleuven.ac.be (A. Van Der Borght). 0733-5210/$ - see front matter q 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcs.2004.09.008

0.50.9%), and carbohydrate (ca. 7.016.0%, on an as-is basis), the latter being mainly starch and lesser amounts of non-starch polysaccharides (NSP) (Roels, 1997). If the isolated gluten is still able to form a network with viscoelastic properties, it is dened as vital gluten, implying that it still has functionality, e.g. in breadmaking. In spite of the increasing importance of wet wheat processing, the literature on the subject is relatively scarce. This survey brings together and reviews the processes for starch-gluten separation from the two major starting materials (wheat our or whole wheat kernels) and provides details on the effects of different factors on the recovery and drying of starch and gluten. First some basic information on the major components involved, i.e. proteins, starch, NSP and lipids is provided.

2. Components involved in the fractionation process 2.1. Proteins Wheat grain contains about 12% proteins (Belitz and Grosch, 1999) which are found in the endosperm. The proteins can be divided in two main groups: the gluten

222

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237

and non-gluten proteins. Non-gluten proteins (ca. 1520% of total wheat protein) consist of albumins (soluble in water) and globulins (insoluble in water, soluble in dilute salt solutions) (Osborne, 1924). These are mainly monomeric proteins with molecular weights (MW) mostly lower than 25,000, although a signicant proportion have MW between 60,000 and 70,000 (Veraverbeke and Delcour, 2002). Gluten proteins (ca. 8085% of total wheat protein), the main storage proteins of wheat, are insoluble in water. Upon hydration and mixing, they form a strong, cohesive, viscoelastic network that allows the wheat our dough to retain yeast fermentation gases and to produce a light, aerated baked product. Gluten proteins can be divided into gliadins and glutenins. Gliadins have MW between 30,000 and 80,000 (Veraverbeke and Delcour, 2002), are singlechained, and are extremely sticky when hydrated. They are rich in proline and glutamine and have a low level of charged amino acids. Intra-chain cystine disulphide bridges are present. Based on electrophoretic mobility at low pH, there are four types of gliadins, i.e. a-, b-, g-, and u-gliadins (Shewry, 2003). In dough formation, the gliadins act as plasticisers, promoting viscous ow and extensibility which are important rheological characteristics of dough. They may associate with one another or the glutenins through hydrophobic interactions and hydrogen bonds. Glutenins are multi-chained and vary in MW from about 80,000 to several million (Hoseney, 1994; Veraverbeke and Delcour, 2002). They apparently impart to dough its property of resistance to extension. The amino acid compositions of glutenins are very similar to those of gliadins, with high levels of glutamine and proline and low levels of charged amino acids. The glutenins are composed of subunits, linked through disulphide bridges. In addition to these inter-chain cystine bonds, glutenins, like gliadins, also contain intra-chain disulphide bridges. The glutenin subunits are released from glutenin by disulphide reducing agents such as b-mercaptoethanol or dithiothreitol. High MW glutenin subunits (MW between 65,000 and 90,000) and low MW glutenin subunits (MW between 30,000 and 60,000) can be distinguished (Goesaert et al., 2004; Veraverbeke and Delcour, 2002). During grain development, wheat storage proteins are deposited as protein bodies which as the grain matures, lose their distinct structure and form a continuous matrix within the endosperm cells, in which starch granules are imbedded. 2.2. Starch Starch is the most abundant component (ca. 6372%) of wheat (Lineback and Rasper, 1988) and is present in the endosperm. It consists of the glucose polymers, amylose and amylopectin. Amylose is essentially linear, consisting of (1/4)-a-linked D-glucopyranosyl units with MW in the range 10 5K10 6 (Lineback and Rasper, 1988). In contrast, amylopectin is highly branched and consists of chains (1/4)-a-linked D-glucopyranosyl units joined

through (1/6)-a-linkages. Branch points occur at approximately every 2025 glucopyranose residues. Amylopectin has one of the highest MW (O108) among naturally occurring polymers. Typical levels of amylose and amylopectin are 2528% and 7275%, respectively (Colonna and on, 1992). Bule Wheat starch granules are of two sizes; large, lenticular A-type granules ca. 1540 mm (mean diameter of ca. 20 mm), and small, spherical B-type granules ca. 110 mm (mean diameter 5 mm) (Karlsson et al., 1983; Lineback and Rasper, 1988; Moon and Giddings, 1993). When viewed in polarised light, native starch granules are birefringent and exhibit a Maltese cross pattern, indicating an orderly arrangement of the starch molecules. The characteristic X-ray pattern and NMR spectra of starch granules is due to the orderly packing of adjacent branches of the amylopectin components. 2.2.1. Damaged starch During milling a small, but signicant proportion (58%) of starch granules in our are physically damaged. Granule fragments produced during milling are not birefringent (Hoseney, 1994). The level of starch damage varies with the severity of grinding and the hardness of the wheat. The rate of water absorption during dough making and enzymic degradation of starch increases with increasing damage. 2.3. Non-starch polysaccharides (NSP) Wheat contains polysaccharides other than starch. NSP are present in the walls of cells of the endosperm and bran tissues which are composed of arabinoxylans (AX) (1/ 3,1/4)-b-glucans, cellulose and arabinogalactan-peptides (AGP). AX are the most abundant (1.52.5% in our) and are water extractable (WE-AX, typically 0.5% in our) or water unextractable (WU-AX, typically 1.5% in our). AX are made up of (1/4)-b-linked D-xylopyranosyl residues, substituted at the C(O)3 and/or the C(O)2 position with a-Larabinofuranosyl units (Izydorczyk and Biliaderis, 1995; Perlin, 1951a,b). The C(O)5 of some arabinofuranosyl units may be esteried with ferulic acid (Fausch et al., 1963; Izydorczyk and Biliaderis, 1995). Wheat arabinogalactan-peptides (AGP) consist of large polysaccharide moieties (9294%) covalently linked to a 15 amino acid peptide (68%) (Van den Bulck et al., 2002). The polysaccharide is built up of highly branched backbone of D-galactopyranosyl residues which are (1/3)-b- or (1/ 6)-b-linked. This galactan backbone is substituted with single arabinofuranosyl residues (Fincher et al., 1974). 2.3.1. Viscosity and water binding capacity WE-AX and solubilised AX (S-AX) (obtained by xylanase hydrolysis of WU-AX or by alkali treatment of WU-AX) yield highly viscous solutions. Their viscosity depends mainly on AX chain length (Dervilly-Pinel et al., 2001; Picout and Ross-Murphy, 2002). WU-AX have

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237

223

a strong tendency to absorb water and swell (Gruppen et al., 1993). Increasing Farinograph absorption upon addition of WU- and WE-AX to our is interpreted as resulting from their water binding capacity (Jelaca and Hlynca, 1971; Kim and DAppolonia, 1977). Endogenous AX have a negative effect on dough handling properties. This may result from immobilising some of the water necessary for complete hydration of gluten proteins and/or due to interference of AX in interactions between the gluten proteins (Roels et al., 1993). 2.4. Lipids Wheat contains about 2% lipids (Belitz and Grosch, 1999). Similar levels of non-polar and polar lipids are present. Based on solubility under specic extraction conditions, they are classied as starch lipids and free and bound, non-starch lipids (Eliasson and Larsson, 1993; Hoseney, 1994). Non-starch lipids comprise about 75% of total our lipids and consist predominantly of triacylglycerols, as well as of other non-polar lipids, and digalactosyl diacylglycerols. The lipids bound to starch (25%) are generally polar. The major constituents are lysophospholipids, in particular lysophosphatidylcholine (lysolecithin). The wheat lipids interact with proteins in dough or gluten systems (Chung, 1986). In fact, most of the free non-starch lipids bind to gluten during dough mixing (Bushuk, 1986; Hoseney, 1994). Lipids that are strongly bound in starch granules are essentially unavailable until starch is gelatinised.

3. Fractionation of our hrmann (1989) list a total of 15 processes Kempf and Ro for the industrial production of wheat starch starting from our or whole kernels, but only a few of these have been used extensively in industry. Fractionation of whole kernels often leads to an impaired gluten production as a result of harsh processing conditions which devitalize the gluten. Also the wet-milling process for extracting starch and gluten directly from wheat kernels is hindered by the high tendency of gluten proteins to agglomerate. This results in the entanglement of gluten proteins with bran components from which the proteins then can no longer be separated (Meuser et al., 1989). Therefore, the most widespread methods for isolation of starch and gluten use wheat our as starting material. 3.1. Overview of the separation methods Several methods have been developed for fractionation of wheat our. Wheat our is essentially endosperm tissue reduced to small particle size by milling, which also separates bran and germ from the endosperm. In most processes, the our in suspended in an aqueous medium

and upon hydration, gluten proteins agglomerate or coagulate and a gluten protein enriched phase or gluten is formed. In this context, we use the term gluten protein agglomeration, rather than the term gluten (protein) coagulation. The gluten protein particles are stabilised by disulphide bonds and secondary bonding forces, such as hydrogen, ionic and other non-covalent bonds. Many laboratory scale experiments use a set of sieves with different pore sizes to recover the gluten from wheat our slurries. In such experiments, the gluten protein agglomeration index (or coagulation index) is calculated from the ratio of gluten on the 400 mm top sieve to the total gluten recovered on a set of sieves. According to Hamer et al. (1989), this parameter measures the degree of gluten protein agglomeration. A special type of gluten protein agglomeration takes place under dough making conditions (high energy input, relatively low water levels). During dough mixing, gluten proteins are stretched and bonds are broken and subsequently, new bonds are formed during dough resting. The result is a gluten matrix (network) that is much stronger than the gluten protein agglomerates formed under more dispersed conditions in our suspensions and with lower energy input. The strong agglomeration during dough making is referred to as gluten development. In most industrial fractionation processes, starch is separated from the protein fraction by centrifugation, by tabling, in hydrocyclones or in decanters, taking advantage of the higher density of starch granules than gluten particles, or by sieving (screening), taking advantage of the difference in size of starch granules and gluten particles (Dahlberg, 1978; Johnston and Fellers, 1971; Knight and Olson, 1984; Verberne and Zwitserloot, 1978). Continuous centrifugation yields a dense bottom stream, called prime starch or A-starch. The top streams carry the gluten fractions and a gelatinous layer, generally referred to as squeegee starch, starch tailings, sludge, low grade starch or B-starch (chereinafter referred to as squeegee starch). This fraction consists of the smaller starch granules, damaged starch, water unextractable NSP and low levels of proteins and ash (Hoseney, 1994; Lineback and Rasper, 1988). The liquid stream contains the water soluble material, referred to as water extractable fraction (WEF). In hydrocyclones, the starch/gluten suspension is fed tangentially into the upper part of the cone. The resulting spinning effect generates the centrifugal force for the separation of the starch granules, which exit from the bottom (apex) of the cone, from the gluten particles and WEF, which exit at the top. In starch tabling, a suspension of starch and gluten ows over an inclined table, which allows the starch to settle and the gluten particles, water and WEF to run off (Kerr, 1950; Watson, 1964; Watson et al., 1951). The surface of the starch deposited on the table is then washed with water and the squeegee starch removed (Watson et al., 1951). Decanters follow the same principle. Sieving or screening retains the agglomerated gluten proteins while the starch milk (also containing the WEF)

224

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237

passes through the sieve. Starch is recovered by centrifuging or tabling the starch milk. Flour fractionation processes can be classied on the basis of the degree of gluten protein agglomeration. The most important processes, i.e. the dough, doughbatter, batter and other batter-based processes are discussed below and an overview given in Fig. 1. Currently, the dough, doughbatter and other batter-based processes (Raisio) are used in North America, while a variation of the Raisio process is popular in Europe (Sayaslan, 2004). Generally, gluten obtained with these processes is vital when dried carefully. In addition, an alkaline process is described. 3.1.1. Dough process The well known Martin or dough(-ball) process is the traditional method for obtaining gluten from wheat our and has been widely used since 1835 (Anderson, 1967). In this process, a stiff dough with approximately 4060% water (Fellers, 1973), is allowed to rest and fully hydrate, producing a centimetre- to metre-sized gluten matrix. The dough is then washed in a continuous kneader with additional water to remove starch and the WEF from the gluten. However, care must be taken to recover the gluten matrix. The starch containing washings (starch milk, starch suspension) are centrifuged, the WEF removed and the lighter squeegee starch separated from the heavier prime starch which is subsequently dried (Anderson, 1967; Fellers,

1973; Rao, 1979). Typical starch yields are between 45 and 60% based on the starting our (Table 1) (i.e. ca. 80% based on the initial starch content in the our) and gluten yields are hrmann, between 10 and 15% (Fellers, 1973; Kempf and Ro 1989; Knight and Olson, 1984; Lineback and Rasper, 1988). 3.1.2. Doughbatter process The rst step in this process is the formation of a stiff dough. The gluten matrix is (partially) dispersed by addition of extra water and mixing, resulting in a batter which is then sieved or centrifuged to obtain starch and gluten. The starch yield as percentage of our is 6979% and total gluten yield is between 11 and 14%. The percentage of our protein recovered as gluten varies between 72 and 91% (Hamer et al., 1989) (Table 1). 3.1.3. Batter process A variation of the Martin process was developed during World War II. This procedure, also known as the batter, slurry or screening process, has been adapted for continuous processing (Lineback and Rasper, 1988). It includes the formation of a slack dough or batter by mixing approximately equal amounts of our and water. The batter is mechanically broken up in the presence of additional water to produce suspended curds of gluten particles containing low levels of residual starch. In this process, millimetresized gluten particles are produced. Typically, temperatures

Fig. 1. Overview of the major processes for separating wheat our into starch, gluten and by-products [squeegee starch (SQ) and water extractable fraction (WEF)].

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237 Table 1 Fractionation of our: yields and protein contents Process Starch Yield (%) Dough Prime SQ Prime SQ Prime Total 4555 1020 59 8 60 6877 6979c 5678c 32 26
a

225

Gluten [Protein] (%) n.a. n.a. 0.25 24 n.a. 3 n.a. !1 0.6 1.6 Yield (%) 1015b 15b [Protein] (%) 7080 80

References

hrmann (1989) Kempf and Ro Knight and Olson (1984) Fellers (1973) Lineback and Rasper (1988) Anderson et al. (1960) Hamer et al. (1989) Johnston and Fellers (1971) Zhuge et al. (1991)

Batter Doughbatter Centrifugation Pre-treatmentd

Prime SQ

713b 1114b 6 13

n.a. n.a. 2040

SQ, squeegee starch; [protein], protein concentration; n.a., not available. a Based on the starting our. b Gluten obtained is vital. c Probably total starch, i.e. prime and squeegee starch. d Pearling prior to milling, followed by dough process.

of 4055 8C are used. The gluten curds are recovered on gyrating sieves while the starch milk and WEF pass through. The starch is recovered by centrifugation or tabling and subsequently dried (Anderson, 1967; Anderson et al., 1958; Fellers, 1973; Knight and Olson, 1984; Rao, 1979). Crude starch recovery in a pilot-plant continuous process based on the starting our is between 68 and 77% with an average of 3% protein (Anderson et al., 1960). Based on the starting our, gluten recovery, on a moisture free basis, is between 7 and 13% (Anderson et al., 1960) (Table 1). The protein recovered in gluten as a percentage of total protein in our is between 72 and 86% (Anderson et al., 1958, 1960) (Table 1). 3.1.4. Other batter-based processes In the Fesca or direct centrifugation process, wheat our is rapidly mixed with water to form a thin batter in which gluten protein agglomeration is minimized by additional shear and a lower slurry temperature (30 8C) compared to the batter process (Fellers, 1973; Johnston and Fellers, 1971). The centrifugal process produces micrometre-sized gluten protein agglomerates. The starch is removed from the suspension by centrifugation while the protein remains suspended (Fellers, 1973; Knight and Olson, 1984; Rao, 1979). The centrifugation yields essentially three layers: prime starch, squeegee starch and a semi-clear supernatant, which contains proteins and WEF. Depending on the our, the starch yield based on the starting our is between 56 and 78% (on moisture free basis) with a protein content of 1% or less in a continuous decanter-type centrifuge (Johnston and Fellers, 1971). The starch as a percentage of total our starch is 6387% (on moisture free basis). The protein concentrate has a protein content of 2040% on a moisture free basis (Johnston and Fellers, 1971; Kerkkonen et al., 1976) and corresponds to 9094% of total our protein (Table 1). The protein fraction can be concentrated by

evaporation and vital gluten recovered (Johnston and Fellers, 1971). The Raisio process is essentially a variation of the Fesca process. Flour and water are combined into a batter which is homogenised to give a uniform suspension without gluten protein agglomeration. Continuous centrifugation of the suspension separates the heavy prime starch from a proteinrich fraction. Protein agglomeration and gluten development then takes place during pin mixing. Gluten is separated by passing over vibrating screens and is then washed, dewatered and dried. The protein content of the gluten is approximately 80%, with a gluten protein recovery of more than 90%. Starch from the gluten screen is almost all squeegee starch, with low levels of prime starch (Dahlberg, 1978; Knight and Olson, 1984; Wadhawan, 1988). A variant of this process is widely used in industry in Europe. Whereas in the Raisio process shear is created by rapidly moving pins through a slow moving slurry, in the variant process a rapidly moving slurry is pumped through a stationary valve (Sayaslan, 2004). 3.1.5. Alkaline process During World War II, an alkaline process was developed. The wheat protein is dispersed and partly dissolved in 0.03 M sodium hydroxide or ammonia. The dispersion is tabled or centrifuged (Anderson, 1967; Anderson et al., 1958; Fellers, 1973; Knight and Olson, 1984) to give a nonvital gluten fraction. A signicant portion (7080%) of total starch is recovered as prime starch (Fellers, 1973). 3.2. Factors affecting the fractionation of our Since most literature discussing the factors affecting our fractionation is based on laboratory dough and doughbatter processes, the following discussion deals mainly with factors affecting dough properties and gluten protein agglomeration. However, large scale commercial

226

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237

production of gluten follows the same principles as those in laboratory procedures. The impact of our composition (protein, starch, NSP and lipids), mixing water (amount, temperature, mineral content), mixing time and speed, washing, processing aids (ascorbic acid, enzymes) and pearling prior to the milling on the separation of gluten and starch are discussed. Table 2 lists the different factors. 3.2.1. Flour The separation behaviour of our differs with the cultivar (Davin et al., 1984; Hamer et al., 1989; Larsson and Eliasson, 1996b; Roels et al., 1998b; Sindic et al., 1993; Weegels et al., 1988; Wootton and Mahdar, 1993). The correlation between breadmaking quality and separation behaviour of a given our is not so clear. In a dough process, ours with good breadmaking properties resulted in high gluten yields (Davin et al., 1984). In a doughbatter process, the baking properties of the our, in particular optimal mixing time and baking absorption, were correlated with gluten yield and gluten recovery (Roels et al., 1998b). However, Hamer et al. (1989) found no correlation between the overall breadmaking performance of the our and gluten yield and the gluten protein agglomeration index. In addition, the processing properties of a given wheat cultivar can signicantly vary as a result of climate and cultivation factors (Prieto et al., 1992; Weegels et al., 1988). This leads to different levels of protein, starch and minor our components (NSP and lipids) which can interfere in the separation of starch and gluten and to different properties, e.g. gluten protein agglomeration index, damaged starch content, of the wheat constituents. The levels and properties of the wheat constituents in our also depend on the milling conditions. For instance, the levels of NSP and damaged starch increase with the extraction level. High-grade/lowextraction ours are much more suitable for production of prime starch and gluten, yielding similar levels (on a wheat basis) of prime starch and higher quality-gluten than lowgrade/high-extraction ours. Indeed, lower grade our increases the squeegee starch levels (Dik et al., 2002). Furthermore, a good correlation exists between the gluten protein agglomeration index and the particle size distribution of the our (as inuenced by milling) (Hamer et al., 1989). Samples with high amounts of small particles showed slow agglomeration (Kelfkens and Hamer, 1991a). 3.2.1.1. Protein. A very important factor in our fractionation is the our protein content. Many researchers have recorded a positive correlation between protein content and gluten yield (Davin et al., 1984; Hamer et al., 1989; Kelfkens and Hamer, 1991a; Prieto et al., 1992; Roels et al., 1998b; Sindic et al., 1993; Wootton and Mahdar, 1993). Logically, starch yields decrease with increasing our protein contents (Kelfkens and Hamer, 1991b). However, Hamer and coworkers (1989) and Roels and coworkers (1998b) did not nd a good correlation between gluten

protein agglomeration index and protein content of the our, although this relationship vary from year to year (Sindic et al., 1993). Additionally, the agglomeration properties of the gluten proteins are of vital importance. In a doughbatter process, Weegels and coworkers (1988) found protein solubility (721%), recovery (5280%) and purity of the protein fraction (6080%) to be strongly dependent on the wheat cultivar. This may have been due to different gluten protein agglomeration capabilities in the different cultivars, which can be related to differences in glutenin aggregation (Roels et al., 1998b). Cultivars with a high degree of glutenin aggregation are characterised by good protein agglomeration, which in turn results in a high gluten yield (Roels et al., 1998b; Weegels et al., 1988) and starch that is less contaminated with protein (Roels et al., 1998b). Both Roels and coworkers (1998b) and Weegels and coworkers (1988) mainly studied the doughbatter process, but obviously gluten protein agglomeration properties may also be signicant in the dough and batter processes. Another important our characteristic is the level of water extractable protein (Weegels et al., 1988). These valuable proteins are present in the waste water and cause problems in its treatment. Among the many water extractable proteins are enzymes, although in general hydrolytic enzyme activity in our is low and their presence is of minor importance. However, in the case of sprout damage, hydrolytic enzyme activity is increased, starch and gluten protein structure can be altered to a signicant extent. Moderately sprout damaged wheat (falling number w150 s) can still be used for the production of starch and gluten with little, if any, effect on gluten yield and quality. However, starch yield and viscosity properties are reduced and the amount of solubles in the waste water increased (Kelfkens and Hamer, 1991a). In addition, proteolytic enzymes are associated with vital wheat gluten (Bleukx et al., 1997) and the wet gluten, if not dried quickly, can be partially hydrolysed. 3.2.1.2. Starch. The starch content in gluten depends strongly on the efciency of gluten washing during isolation, indicating that starch is physically entrapped in the gluten network (Saulnier et al., 1997). Indeed, starch has been suggested to act as inert ller in dough making (Bloksma, 1990). There is evidence that starchgluten interactions in dough and starch play an active role in determining dough rheological characteristics (Larsson and Eliasson, 1997; Petrofsky and Hoseney, 1995). However, the nature of the interactions are unclear. Soft wheat doughs had higher storage and loss moduli than hard wheat starch doughs, possibly because of more intense starchgluten interactions. Seguchi and coworkers (1998) concluded that aging negatively impacts the separation of our into starch and protein fractions. In particular, the fraction of dry matter recovered in the squeegee starch increases steadily with aging. They speculated that the main binding forces between

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237 Table 2 Factors affecting the fractionation of our Parameter Flour Authors Larsson and Eliasson (1996b), Roels et al. (1998b), Sindic et al. (1993), Weegels et al. (1988) and Wootton and Mahdar (1993) Weegels et al. (1988) Dik et al. (2002) Weegels et al. (1988) Roels et al. (1988b) and Weegels et al. (1988) Hamer et al. (1989) Hamer et al. (1989), Kelfkens and Hamer (1991a) and Wootton and Mahdar (1993) Weegels et al. (1988) Petrofsky and Hoseney (1995) Seguchi (1993) and Seguchi et al. (1998) Christophersen at al. (1997), Frederix (2003), Roels et al. (1993, 1998a) and Wang et al. (2003, 2004) ke s et al. (1983a, b), Bushuk (1986), Roels et al. (1998a) and Be Zawistowska et al. (1985) Larsson and Eliasson (1996b) Dik et al. (2002) Larsson and Eliasson (1996a,b) and Robertson and Cao (1998) Frederix et al. (2004a) Anderson et al. (1960) Fellers (1973) Temperature Anderson et al. (1960) lu et al. (2002) ndem-Makasciog Yo Ionic strength Mixing speed and mixing time Ascorbic acid Larsson (2002) Frederix et al. (2004a) Findings Separation behaviour depends on cultivar Processing properties vary with climatological and cultivation factors Low-extraction ours are more suitable for separation Gluten agglomeration capabilities depend strongly on the wheat type Good gluten agglomeration leads to a better separation Coagulation index is an indicator for the rate of gluten agglomeration Positive correlation between protein content and gluten yield

227

Protein

Starch

NSP Lipids

High level of water extractable protein represents a loss of valuable raw material Evidence for starchgluten interactions Aging negatively impacts the separation of our due to increased hydrophobic interactions NSP have a negative effect on the separation Lipids and gluten proteins interact Lipids have a negative effect on glutenstarch separation Lipid hydrolysis may be a reason for the difcult separation with aging Increased separation with increasing water-to-our ratio in the dough process Increased separation with increasing water-to-our ratio in the dough batter process Increased separation with decreasing water-to-our ratio in the batter process Inadequate water levels have a negative impact on separation in the dilute batter process (Fesca) Increasing temperature positively impacts recovery of protein in gluten in the batter process Decreasing temperature leads to more prime starch and less protein in gluten in the doughbatter process Sodium chloride reduces the level of starch and gluten and yields more liquid and unseparated dough Increasing mixing time and speed improve gluten agglomeration

Water content

Anderson et al. (1960) Endo et al. (1989) Larsson and Eliasson (1996b) Dik et al. (2002) Hamer et al. (1989) Christophersen et al. (1997)

Optimum recovery occurs under conditions requiring greater energy input Ascorbic acid enhances the separation of gluten and starch Ascorbic acid improves separations Ascorbic acid does not improve or even worsens separations Enzymes hydrolysing NSP are powerful tools to improve the processing properties of poor quality wheats Selectivity of xylanases towards soluble and insoluble arabinoxylan is of major importance for improving separation; decreased viscosity goes hand in hand with good separation Transglutaminase improves separation Cellulase and protease preparations improve protein recovery and coagulation index of a our with moderate processing properties. Protease and amylase increase the viscosity of the starch slurry, resulting in a less efcient separation of starch and solubles

Enzymes

Olson (2002) Weegels et al. (1992)

prime and squeegee starches in stored our were hydrophobic. Indeed, in a previous study Seguchi (1993) showed that hydrophobicity of the starch granules increased with aging. Similar increases were observed when the ours were chlorinated or heat treated for short times at higher temperatures.

Another important our characteristic is the level of damaged starch which depends on the severity of grinding and hardness of the wheat (Hoseney, 1994). Damaged starch granules are part of squeegee starch and reduce prime starch yield. In addition, damaged starch can absorb three times more water than native starch and can, therefore, interfere

228

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237

with gluten protein agglomeration (Section 3.2.2.1). The presence of damaged starch in the prime starch may restrict the use of wheat starch in certain applications. 3.2.1.3. Non-starch polysaccharides. Overall dough characteristics are negatively correlated with total AX (Rouau et al., 1994). NSP and in particular AX have a negative effect on gluten protein agglomeration during starch-gluten separation (Christophersen et al., 1997; Frederix et al., 2002, 2003; Roels et al., 1993, 1998a; Wang et al., 2002, 2003). Due to their high water binding/holding capacity, AX may interfere indirectly with gluten formation by competing for water and thus changing conditions for gluten protein agglomeration. This can play an important role in processes where a dough making step, i.e. a stage with a limited amount of water, is included, e.g. in the Martin and the doughbatter processes. In addition, the amount of water retained in the gluten during isolation is largely determined by AX. Hence, the level of NSP associated with the gluten determines the energy required to dry the isolated gluten (Roels et al., 1998a). Furthermore, a clear distinction should be made between the impact of WE-AX and WU-AX. The negative effect of WE-AX in batter or doughbatter processes can be attributed to a large extent to their impact on batter viscosity. Reduced viscosities of wheat our slurries go hand-in-hand with good gluten protein agglomeration because the mobility of the components increases (Christophersen et al., 1997; Frederix et al., 2003; Redgwell et al., 2001). According to Redgwell and coworkers (2001), the batter viscosity can be attributed mainly to the size and the level of WE-AX or may in part be dictated by non-covalent interactions either between AX themselves or AX and other macromolecules such as starch and gluten proteins. Steric hindrance of high MW AX for gluten protein interactions can be responsible for the negative effect of these polysaccharides on gluten protein agglomeration (Weegels et al., 1990). Wang and coworkers (2002, 2003) suggest that the WE-AX bound ferulic acid moieties can act in two direct ways, rstly, by cross-linking WE-AX, leading to a higher viscosity in the liquid phase and secondly, by cross-linking WE-AX to gluten proteins. WU-AX, which are present in discrete cell wall fragments, may interfere with gluten protein agglomeration by forming a physical barrier against gluten protein interactions (Frederix et al., 2002, 2003; Wang et al., 2002, 2003). In addition to the water holding capacity and the particulate nature of WU-AX, Wang and coworkers (2002, 2003) also ascribed the negative effect of WU-AX on gluten protein interactions to a ferulic acid mediated reaction. Finally, higher proportions of AGP are present in gluten fractions exhibiting the best agglomeration properties (Roels et al., 1998a) suggesting that AGP may play a signicant, but as yet not understood role in wheat gluten protein agglomeration.

3.2.1.4. Lipids. Interactions between lipids and gluten proteins has been demonstrated. They occur during gluten ke s et al., protein agglomeration and dough mixing (Be 1983a, b; Bushuk, 1986; Olcott and Mecham, 1947; Roels et al., 1998a; Zawistowska et al., 1985). In the presence of lipids, gliadin, as well as trace amounts of albumin and globulin proteins, can form aggregates of high apparent ke s et al., 1983a). The lipid, mainly galactolipid, MW (Be appears to play a key role in the formation of these ke s et al., 1983b). Indeed, gliadins containing aggregates (Be associated galactolipids are preferentially incorporated into the gluten protein agglomerates (Roels et al., 1998a; Zawistowska et al., 1985). Ultracentrifugation of dough made from chloroform defatted our improves separation of the gluten and starch and may indicate that in certain conditions lipids (non-polar or polar) have a negative effect on glutenstarch separation (Larsson and Eliasson, 1996b). During aging, changes in our lipids occur, including slow hydrolysis leading to formation of free fatty acids (Dik et al., 2002). This may be one of the reasons for the difculty of separation of aged our into starch and gluten (Seguchi et al., 1998). 3.2.2. Water In the following sections the inuence of the level, temperature and mineral content of mixing water on dough properties and gluten agglomeration and subsequent fractionation are discussed. 3.2.2.1. Water content. Since each process uses a different level of water (Section 3.1), the different processes will be discussed separately. Dough process. The separation efciency in the dough process depends on the dough water content. The optimum water-to-our ratio for dough development varies with wheat our (Knight and Olson, 1984). Larsson and Eliasson (1996a) prepared doughs with water contents varying between 38 and 48.6% and subjected the doughs to high centrifugal forces (100 000g). At the highest water content (48.6%), good separation into a starch phase, a gluten phase, a squeegee layer and a liquid phase was obtained. With decreasing water content (starting from a water content of 47.1%), the separation efciency decreased and increasing levels of unseparated dough appeared below the starch phase, probably due to the reduced availability of water for initial gluten hydration and development during dough mixing. At a water content corresponding to the Farinograph water absorption (500 BU, 45.3%), the separation was still acceptable. At the lowest water content (3840%), no liquid phase was observed and the fractionation was strongly hindered or did not occur at all. Thus, a minimum level of water was needed to separate gluten and starch (Larsson and Eliasson, 1996a,b). This is supported by the data of Robertson and Cao (1998a,b) who found with a water-to-our ratio between 0.6 and 0.95 (corresponding to a water content of 4555%) the separation of the our into

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237

229

gluten and starch improved as this ratio increased. However, in the range 0.951.1 (corresponding to a water content of 5558%), the separation efciency decreased and protein was lost, suggesting a decreased matrix cohesiveness. Doughbatter process. As in the dough process, there is positive leffect of increased water levels on gluten protein agglomeration in the dough making step of the doughbatter process (Frederix et al., 2004a). This again may be explained by the availability of more water for initial gluten protein hydration and gluten development during dough mixing. Batter process. Anderson and coworkers (1960) studied the batter process the effect of the water-to-our ratio in. They mixed water and our in ratios between 1 and 1.8 and washed the resulting batter with additional water. Gluten recoveries increased with decreasing water-to-our ratios. Schurgers et al. (2004) also found the gluten protein agglomeration index, gluten protein recovery and gluten yield increased with decreasing water content. The optimum water-to-our ratio varied with the type of wheat our (Knight and Olson, 1984). Fesca process. In the Fesca process, the water-to-our ratio is critical. Too much or insufcient water has a negative impact on separation. With too much water, gluten settles on top of the starch to yield a sticky sludge layer. With too little water, the slurry develops into a dough that does not separate at relatively low centrifugal forces. The optimum water-to-our ratio depends on the wheat our (Knight and Olson, 1984). Hard wheat, high protein ours require as much as 1.8 parts of water per part of our while soft wheat, low protein ours require as little as 1 part of water per part of our (Fellers, 1973; Johnston and Fellers, 1971). 3.2.2.2. Temperature of mixing water. Dough process. Optimum gluten development is generally obtained within a mixing temperature range of 24.528 8C. Higher temperatures favour contacts and reactions among the gluten proteins, hence gluten development takes place more quickly (Veal, 1954). lu and co ndem-Makasciog Doughbatter process. Yo workers (2002) studied the use of lukewarm water (20 50 8C) for separation in the doughbatter process as a means of strengthening the gluten matrix. As the temperature decreased, the fraction of dry matter in the squeegee starch decreased in favour of that in the prime starch fraction. However, as the temperature increased, protein recovery in the protein fraction increased slightly and the washing/sieving process was faster. The latter is probably due to the lower viscosity at the higher temperatures. Experiments at 50 8C yielded a dough which was extensively dispersed, sometimes to the point of being able to pass through the sieve openings, resulting in little, if any, protein recovery. Furthermore, when the washing temperature differed from the dough making temperature, the doughs tended to disintegrate.

Batter process. In the batter process, the temperature of the mixing water inuences the recovery of the our protein in the gluten (Anderson et al., 1960). Typically, temperatures of 4055 8C are used. These high temperatures facilitate gluten protein hydration and agglomeration and reduce the mixing time necessary to produce a batter of suitable consistency (Anderson et al., 1958). The consequences of such temperatures on gluten functionality have not been reported. Fesca process. The optimum slurry temperature in the Fesca process, w30 8C, is related to the reduced gluten protein agglomeration observed in this process. This is in contrast to the batter process, where temperatures of 40 55 8C are used to enhance gluten protein agglomeration (Fellers, 1973; Johnston and Fellers, 1971). 3.2.2.3. Mineral content of mixing water. Mineral salts in doughs usually strengthens dough properties by promoting glutenin and gliadin aggregation (Kim and Bushuk, 1995; Weegels et al., 1990). Similarly water used for dough development usually contains some mineral salts. Water of low salt content may lead to slimy gluten (Knight and Olson, 1984). Gluten proteins have only few charged groups, and, at low ionic strength, salt ions may decrease repulsion and enhance aggregation (salting out) (Weegels et al., 1990; Wellner et al., 2003). Salt concentrations ranging from 0 to 5% (on our weight) increase the mixing time necessary for optimal dough development (Hlynka, 1962). Washing gluten from doughs with a 2% sodium chloride solution instead of distilled water increased nitrogen recovery in the gluten and decreased its lipid content (Mecham and Weinstein, 1952). Glutens isolated with wash water containing increasing levels of sodium chloride had increasingly higher nitrogen contents (Fu and Sapirstein, 1996). These results show that salt enhances gluten protein agglomeration and therefore can improve starchgluten separation, in particular gluten yield and purity. Furthermore, addition of salt to the washing step in the dough process has an additional benet by decreasing binding of lipid to the gluten proteins and thereby extending gluten shelf-life. However, ultracentrifugal separation of doughs with added sodium chloride (0.5 and 1% on our weight) is inferior to doughs with water alone. Sodium chloride reduced the yields of starch and gluten and more liquid and unseparated dough was obtained after centrifugation (100 000g, 1 h) (Larsson, 2002). 3.2.3. Mixing and washing 3.2.3.1. Mixing time and speed. In the doughbatter process, increased mixing times and speeds during dough development improve gluten protein agglomeration because the gluten network is more developed (Frederix et al., 2004a). Anderson and coworkers (1960) also reported

230

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237

optimum recovery in the batter process under conditions of greater energy input (i.e. longer mixing time and/or higher mixing speed). Increasing mixing time leads to a more efcient extraction of starch and separation is favoured. However, overmixing must be avoided, because the gluten matrix becomes weaker, its MW is reduced and protein solubility increases (Larsson and Eliasson, 1996b; Robertson and Cao, 1998a,b). These observations would most likely also apply to the dough process. 3.2.3.2. Washing process. The severity and nature of the washing process also affects starch and gluten recovery. Successful washing depends on the ability to control and maintain or strengthen proteinprotein interactions and weaken proteinstarch interactions. Protein interactions depend on the level of hydration, physical work, rest or relaxation periods, and the severity and nature of the washing operation. In general, proteinstarch separators use large amounts of water, remove water-soluble proteins, and discharge large amounts of dilute protein-bearing aqueous waste. In addition, separators and the associated dryers use excessive energy. In a laboratory based dough washing experiment, Robertson and Cao (1998a,b) substituted washing water with concentrated ethanol. A potential advantage of an ethanol-washing process would be a reduction of the energy required for gluten drying. Such a medium may, however, remove lipid-associated proteins and gliadins. In addition, the displacement of water from the hydrated structure may lead to progressive fragmentation of the gluten structure and, if continued indenitely, to its total displacement and, nally, to a dry, inseparable powder (Robertson and Cao, 1998a). Thus the physical form of the gluten produced by ethanol washing will differ from that obtained by water washing. Generally, the ethanol washed gluten has low cohesiveness and breaks into smaller clumps than the water-washed gluten. The gluten from ethanol washing is also less sticky and dries more rapidly than water-washed gluten. The yield of solvent-soluble whole gluten strongly depends on ethanol concentration and temperature. Efcient separations are obtained with an ethanol temperature of K13 8C and a water content of 40%, v/v (60%, v/v, ethanol). The effectiveness of the low temperature was attributed to the reduced solubility of gliadin proteins and lipoprotein complexes (Robertson et al., 1999). 3.2.4. Other processing aids 3.2.4.1. Ascorbic acid. Ascorbic acid enhances the separation of gluten and starch in the dough and batter process (Endo et al., 1989). However, Dik and coworkers (2002) concluded that ascorbic acid does not improve the separation behaviour in the doughbatter process and may even cause inferior separations. The amount of dry matter (ca. 9%) lost from the protein fraction on addition of ascorbic acid appeared mainly in the squeegee starch.

Although gluten levels are lower, ascorbic acid addition yields a purer protein fraction (Dik et al., 2002). In phase separation of wheat our doughs at high centrifugal forces (100,000g), (Larsson and Eliasson, 1996b) showed that addition of 230 ppm ascorbic acid to the our gave less unseparated dough. Although Larsson (1997) attributed the lower level of unseparated dough on the impact of the ascorbic acid on the interaction between the gluten and starch phases, this may not necessarily be the case as the level of separated gluten was not inuenced by ascorbic acid treatment. 3.2.4.2. Enzymes. Enzymes as processing aids in the separation of wheat our into starch and gluten, and in particular preparations containing hydrolases acting on nonstarchy polysaccharides, especially xylanases, have been investigated. In the doughbatter process, the addition of these partially puried hydrolases preparations boosted the gluten yield of a wheat variety with very poor processing properties (Hamer et al., 1989). This indicated that these hydrolase preparations are powerful tools for improving the processing properties of poor quality wheats. Weegels and coworkers (1992) also found the protein recovery and the gluten protein agglomeration index of our with intermediate processing properties increased signicantly when such an enzyme preparation was added in a doughbatter process. Kelfkens and Hamer (1991a) suggested that the improved gluten protein agglomeration from moderately sprout damaged wheat could be attributed to the increased activity of hydrolases acting on non-starchy polysaccharides in the germinating kernels. In pilot scale, batter process tests with a xylanase from Aspergillus aculeatus (XAA) an improved gluten protein agglomeration and gluten yield and an improved prime starch yield was obtained without affecting the purity of the prime starch. Enzyme addition appeared to have no negative effect on the gluten quality (Christophersen et al., 1997). (High MW) WE-AX and S-AX were rapidly degraded to lower MW fragments and WU-AX was also solubilised to a limited extent. It was concluded that for optimal the xylanase preparation should have high activity towards WE-AX results a low activity towards WU-AX and, in addition, that the solublised AX should be of low MW, (Christophersen et al., 1997). These results suggest important contribution of AX to batter viscosity in glutenprotein interactions. The benecial effect of xylanases has also been ascribed to a decreased interaction between gluten proteins and WE-AX (Redgwell et al., 2001) or removal of the steric hindrance that AX exert on gluten protein agglomeration (Weegels et al., 1990). Frederix and coworkers (2003, 2004b) compared the use of XAA, with selectivity for hydrolysis of WE-AX, to that of Bacillus subtilis xylanase (XBS), with selectivity for hydrolysis of WU-AX, on the wheat our glutenstarch separation in the doughbatter process. In agreement with

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237

231

Christophersen and coworkers (1997), XAA improved the gluten protein agglomeration. In the lower dosage range, XBS had a clearly negative effect on gluten protein agglomeration. This was explained on the basis of increased viscosity of the batter on solubilisation of high levels of WU-AX and the concomitant reduction in mobility of the gluten proteins in the agglomeration process. Higher XBS dosages further degraded S-AX and decreased batter viscosity, allowing improved gluten protein agglomeration behaviour. Frederix and coworkers (2004b) suggested that high dosages of a xylanase able to both solubilise WU-AX and degrade the S-AX and WE-AX would substantially decrease batter viscosity and, hence, improve gluten protein agglomeration. The addition of an XBS mutant, insensitive to the wheat endogenous xylanase inhibitors, improved gluten protein agglomeration at much lower dosages than the wild-type (inhibitor sensitive) XBS. This clearly shows the impact of xylanase inhibitors on xylanase functionality in wheat our fractionation (Frederix et al., 2004b). Furthermore, addition of xylanases at different phases of the doughbatter process indicated that the viscosity during the earlier phases of the batter mixing rather than at the end of the process determines gluten protein agglomeration properties (Frederix et al., 2003). Weegels and coworkers (1992) studied the effect of partially puried cellulase, lipase and protease preparations which contained substantial alpha -amylase activity in the batter mixing step of the doughbatter process on the separation of starch and gluten. The protein recovery and the gluten protein agglomeration index of a our with intermediate processing properties increased signicantly when the cellulase and protease preparations were used. However, the protease preparation resulted in loss of proteins to the soluble fraction. The action of the protease and alpha-amylase increased the viscosity of the starch slurry, because more WEF was released, resulting in a less efcient separation of starch and soluble fraction. Transglutaminases which build new inter- and intramolecular bonds between and within proteins may inuence gluten protein agglomeration. Olsen (2002) has reported that transglutaminase action can, in general, improve the separation of wheat ours in and the purity of particular fractions in the doughbatter process. 3.2.5. Pearling Zhuge and coworkers (1991) compared ground pearled wheat with conventionally milled our (70% extraction) as starting material. Hard red winter wheat was pearled to remove 11.4% of bran and germ and the pearled wheat ground to an average particle size of 200 mm. Optimum yields and purity of gluten and starch were obtained with this average particle size. Indeed, coarse grinding resulted in large endosperm particles, which led to an inefcient gluten development in the dough mixing step, while ne grinding

resulted in severe starch damage. The level of water needed to form a dough was 35% higher for ground pearled wheat than for our. Ground pearled wheat resulted in higher gluten yield than our with the same protein content (13.2% compared to 10.7%). No signicant difference was noted in the yield and purity of prime starch (32.1 with a protein content of 0.6%). The yield of squeegee starch was 25.8% in ground pearled wheat (1.6% protein) and 21.6% in our (0.9% protein). The ground pearled wheat contained 7.3% less damaged starch than the straight-grade our and was 1.2% higher in protein.

4. Whole wheat wet fractionation process The production of starch from our has two important disadvantages when compared to that starting from the whole kernel. Firstly, our extraction in the mill does not exceed 7580%, thus not more than 8590% of the starch originally present in the wheat is available for starch production. Secondly, part of the starch is damaged by the dry milling process. The presence of damaged starch may restrict the use of starch in certain applications and reduces the yields of prime starch. 4.1. Overview of wet separation methods Most of the processes involve the same basic operations, i.e. steeping, milling of the wet wheat and then separation of the bran, germ, gluten (protein fraction) and starch. Fig. 2 provides an overview of the different processes. Industrially, these processes are generally less important. 4.1.1. Halle fermentation An early method for producing wheat starch from whole kernels is the Halle fermentation. In this process wheat is steeped in water for about a week at 25 8C to soften the grain, followed by grinding and fermentation of the mash. Protein is degraded or solubilised. Starch is washed out through the openings of a rotating perforated drums, while the protein and bran fractions remain inside the drum hrmann, 1989). Starch is recovered by (Kempf and Ro tabling or centrifugation, rened, and dried (Fellers, 1973; hrmann, 1989; Knight and Olson, 1984). Kempf and Ro Yields of 60% of total starch, low in protein, are obtained hrmann, 1989) (Table 3), (Fellers, 1973; Kempf and Ro however, vital gluten cannot be recovered (Kempf and hrmann, 1989). Ro 4.1.2. Alsatian process In the Alsatian process, a non-fermentative steeping (1 2 d at 3035 8C) is conducted (Fellers, 1973; Kempf and hrmann, 1989; Knight and Olson, 1984). After steeping Ro the wheat is crushed and the doughy mash washed in a perforated trough with rotating arms. Starch and the wash water pass through the perforations and the starch is

232

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237

Fig. 2. Overview of the major wet separation methods for separating whole wheat into starch, gluten and by-products [bran, squeegee starch (SQ) and water extractable fraction (WEF)].

subsequently tabled or centrifuged, further rened, and dried (Fellers, 1973). A yield of about 70% of total starch was reported by Fellers (1973), whereas Kempf and hrmann (1989) reported yields of about 45% prime Ro starch and 1015% squeegee starch (Table 3). Bran can be separated from the mash by continued washing and about 30 or 40% of the gluten proteins recovered (Fellers, 1973). However, separation of the gluten presents considerable difculty and rarely results in yields of more than 5% of the hrmann, 1989). wheat mass (Kempf and Ro 4.1.3. LongfordSlotter process The LongfordSlotter process is a modied version of the Alsatian process and closely corresponds to the industrial processes for corn starch production. In this wet-milling process, wheat is steeped for 24 h at 37 8C at a relatively high concentration (between 0.3 and 0.5%) of hrmann, sulphur dioxide (Fellers, 1973; Kempf and Ro 1989) to lower pH, increase gluten solubility, and inhibit microbial growth. In addition, sulphur dioxide depolymerises the gluten proteins and loosens the starch/gluten matrix. The steeped wheat is ground and the resulting slurry sieved. The residue comprises the bran fraction. The starch and gluten suspensions are tabled (Fellers, 1973). Starch yields between 55 and 60% with a protein content hrmann, 1989; of 0.2% are obtained (Kempf and Ro

Knight and Olson, 1984). According to Fellers (1973), the yield is about 75% of the total starch in the wheat (Table 3). Since the wheat is initially steeped in the presence of sulphur dioxide, the gluten obtained is not vital. 4.1.4. Pillsbury hydromilling The Pillsbury hydromilling process was developed to obtain maximum endosperm recovery (Rodgers and Gidlow, 1974). Wheat is steeped in aqueous acid at pH 0.81.7 at 3740 8C for 1224 h (Knight and Olson, 1984). The steeping operation is followed by wet grinding and sieving. Germ and bran are retained and removed. The endosperm is dewatered and dried to form a oury product hrmann, 1989). A recovery of about 80% (Kempf and Ro wheat solids is obtained, which corresponds to a 73% yield of straight white our by dry milling to comparable endosperm purities (Knight and Olson, 1984) (Table 3). Starch and gluten can be isolated by making a dough from the our followed by a conventional dough (Martin) or batter process (Knight and Olson, 1984). 4.1.5. Far-Mar-Co process The Far-Mar-Co wet process is similar to the dough or batter process, except that endosperm is used for making the dough or batter instead of dry milled our. Whole wheat kernels are tempered in water and then ake milled

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237 Table 3 Whole wheat wet fractionation process: yields and protein contents Process Starch Yield (%) Halle fermentation Alsatian process Total Prime SQ Total 60 45 1015 70 5060a 75 80 n.a. [Protein] (%) Low n.a. n.a. n.a. 0.2 n.a. n.a. n.a. Gluten Yield (%) !5 [Protein] (%) n.a. References

233

Fellers (1973) and Kempf and hrmann (1989) Ro hrmann (1989) Kempf and Ro Fellers (1973) hrmann (1989) Kempf and Ro and Knight and Olson (1984) Fellers (1973) Knight and Olson (1984) Rao and Shoup (1978)

LongfordSlotter Total Endosperm

Pillsbury hydromilling Far-Mar-Co

n.a.

80

SQ, squeegee starch; [protein], protein concentration; n.a., not available. a Probably prime starch.

(Rao and Gerrish, 1975), pin milled (Rao et al., 1976) or ake milled followed by ake disintegration, to reduce the particle sizes of the kernel components (Rao and Shoup, 1978). The milled particles are mixed with water to form a relatively dilute homogeneous slurry, in which the solid particles are uniformly dispersed in the liquid to obtain a maximum dissociation of the starch and gluten matrix. The dispersion is separated into its gluten, starch and water phases by centrifugation (Rao and Shoup, 1978). The gluten phase is further processed to remove remaining impurities. Vital gluten is recovered with a protein content of about 80% by weight, on a dry basis (Rao and Shoup, 1978) (Table 3). 4.2. Factors affecting the whole wheat wet fractionation process The main factors affecting the wet separation are related to the steeping conditions. 4.2.1. Steeping At room temperature and atmospheric pressure, it takes 1924 h to soften the wheat sufciently for further processing. Steeping time can be reduced considerably by steeping under pressure (Meuser et al., 1989) or by breaking the wet grains between smooth rolls prior to steeping (Kema et al., 1996). With short steeping times of a few hours, the use of sulphur dioxide to inhibit microbial growth is unnecessary and the proteins remain vital. Yuan et al. (1998) studied the effect of steeping conditions (sulphur dioxide, lactic acid, and hydrochloric acid) on wet-milling characteristics of hard red winter wheat. Sulphur dioxide treatments gave the highest starch yield, the lowest average protein content in starch and the highest protein content in gluten, but the lowest gluten yield. Steeping with sulphur dioxide was considerably more effective in dispersing the wheat gluten than the other treatments. In contrast, lactic acid steeping gave the lowest starch yield and the highest average protein content in the starch.

4.2.2. Processing aids The use of enzymes has been suggested for the steeping step in the wet-milling processes. Vaara and coworkers (1989) suggested the application of phytin degrading enzymes in the steeping process, because phytic acid may cause various problems in processing and is nutritionally detrimental. Scott (1991) proposed a method for wet-milling grain in combination with a cellulase containing preparation to produce starch. Olsen and Nielsen (2002) described a process in which cracked, steeped kernels are treated with preparations containing acidic protease, xylanase, cellulase and/or arabinofuranosidase.

5. Drying procedures 5.1. Drying of gluten The nal drying stage to a large extent determines gluten quality. Gluten deteriorates rapidly when kept in a wet state as a result of endogenous or exogenous proteolytic action leading to extremely soft and extensible gluten, which is unsuitable for most of the major uses which rely on retention of the unique cohesive, viscoelastic properties (Redman, 1971). On an industrial scale, production of dried vital gluten presents many difculties because of its sensitivity to drying. Most commercial gluten is dried in a ash or ring drier. Freeze-drying, which produces gluten of the highest vitality, is uneconomic. Commercial gluten is dried to about 610% moisture. The most important methods for drying in commercial gluten production are discussed below. 5.1.1. Flash drying Hydrated gluten is very sensitive to heat and great care has to be taken to avoid devitalisation, while at lower moisture content, gluten is less susceptible to heat denaturation. In ash drying, denaturation is limited by rst blending the wet gluten with a large quantity of previously dried gluten (Wadhawan, 1988). The process

234

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237

consists of the following steps; extrusion, cutting the extruded gluten into short pieces about 10 mm in length and back-mixing dry gluten onto the wet gluten pieces to produce semi-dry fragments at about 2530% moisture (Wadhawan, 1988). The fresh wet gluten is further dried while moving in a stream of warm air (Bushuk and Wadhawan, 1989). 5.1.2. Vacuum drying Vacuum drying was one of the rst methods employed in commercial gluten production. The wet gluten is cut into small pieces or extruded into short fragments and these are dried in a steam heated vacuum ovens. Although the gluten is not exposed to high temperatures, it is seldom used industrially because of the high cost. 5.1.3. Spray drying Spray drying is in use in commercial plants around the world, but is not as common as ash drying. The wet gluten, dispersed in liquid, is sprayed into a tower or chamber with heated air. Several agents, such as ammonia, carbon dioxide and acetic acid have been used as gluten dispersants in spray drying. 5.2. Drying of starch The rst step of drying starch is dewatering of the starch slurry, i.e. removing free water from the starch slurry without applying heat. A concentrated starch slurry is fed to the dewatering equipment, usually a horizontal centrifuge basket equipped with a perforated bowl lined with a lter cloth (Sriroth et al., 1999). Alternatively vacuum or pressure lters which operate at low (vacuum) and high pressure using special lter cloth material can be used (Sriroth et al., 1999). Residual moisture is removed by drying the dewatered starch cake in a stream of hot air. Another popular starch dryer employs a moving belt of metal fabric that carries the starch cake through chambers of circulated air heated to the desired temperature and nally through a cooling chamber (Watson, 1967). The most rapid starch drying is provided by a ash dryer. The starch particles are dried instantly and are collected in dust cyclones (Watson, 1967). However, wheat starch, particularly the smallgranule starch, can form a very explosive mixture (Hoseney, 1994). Most commercial starches are dried to 1215%.

and properties of the our constituents, including gluten protein agglomeration properties, protein and AX content and the level of WE protein, WE-AX and WU-AX (Table 2). Different processes are affected to varying extents by these properties. Process conditions, such as energy input (i.e. mixing time and speed) and the temperature and composition of the mixing water, also have a signicant effect on the separation efciency. Flour separation can be inuenced by various processing aids, such as ascorbic acid and enzymes, in particular xylanases. In this respect, the selectivity of xylanases and their sensitivity to inhibition by the endogenous xylanase inhibitors in wheat to a large extent determine their functionality, particularly in batter and doughbatter systems. In whole wheat fractionation, often harsh process conditions are used, resulting in little if any vital gluten. In these processes, the steeping conditions have the largest impact on the efciency of the separation. The use of enzymes as processing aids the separation has been suggested.

Acknowledgements We thank Christophe M. Courtin and Soe A. Frederix for critical discussions.

References
Anderson, R.A., 1967. Manufacture of wheat starch. In: Whistler, R.L., Paschall, E.F. (Eds.), Starch, Chemistry and Technology. Industrial Aspects, vol. II. Academic Press, New York, pp. 5363. Anderson, R.A., Pfeifer, V.F., Lancaster, E.B., 1958. Continuous batter process for separating gluten form wheat our. Cereal Chemistry 35, 449457. Anderson, R.A., Pfeifer, V.F., Lancaster, E.B., Vojnovich, C., Grifn, E.L., 1960. Pilot-plant studies on the continuous batter process to recover gluten from wheat our. Cereal Chemistry 37, 180188. ke s, F., Zawistowska, U., Bushuk, W., 1983a. Proteinlipid complexes in Be the gliadin fraction. Cereal Chemistry 60, 371378. ke s, F., Zawistowska, U., Bushuk, W., 1983b. Lipid-mediated aggregaBe tion of gliadin. Cereal Chemistry 60, 379380. Belitz, H.-D., Grosch, W., 1999. Food Chemistry. Springer, Berlin. Bleukx, W., Roels, S.P., Delcour, J.A., 1997. On the presence and activities of proteolytic enzymes in vital wheat gluten. Journal of Cereal Science 26, 183193. Bloksma, A.H., 1990. Dough structure, dough rheology, and baking quality. Cereal Foods World 35, 237244. Bushuk, W., 1986. Proteinlipid and proteinlcarbohydrate interactions in ourlwater mixtures, In: Blanshard, J.M.V., Frazier, P.J., Galliard, T. (Eds.), Chemistry and Physics of Baking. The Royal Society of Chemistry, London, pp. 147154. Bushuk, W., Wadhawan, C., 1989. Wheat gulten is good not only for breadmaking. In: Pomeranz, Y. (Ed.), Wheat is unique. Structure, composition, processing, end-use properties, and products. AACC, St Paul, MN, pp. 263275. Christophersen, C., Andersen, E., Jakobsen, T.S., Wagner, P., 1997. rke 49, 512. Xylanases in wheat separation. Starch-Sta Chung, O.K., 1986. Lipidprotein interactions in wheat our, dough, gluten, and protein fractions. Cereal Foods World 31, 242256.

6. Overview Flour and whole wheat kernels are the two logical feed stocks for production of starch and gluten from wheat. Flour has two major disadvantages: milling results in some starch damage and starch is lost in milling side streams. The main our fractionation processes are the dough (Martin), dough batter, batter and dilute batter process. The efciency of the starch-gluten separation is determined by the levels

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237 on, A., 1992. New insights on starch structure and Colonna, P., Bule properties, in: Cereal Chemistry and Technology: a Long Past and a Bright Future. Proceedings of the Ninth International Cereal and Bread Congress, pp. 2542. Dahlberg, B.I., 1978. A new process for the industrial production of wheat rke 30, 812. starch and wheat gluten. Starch-Sta Davin, A., Godon, B., Nicolas, M., Popineau, Y., 1984. Incidences tales sur le rendement dextraction et les caracte ristiques varie . Industries des Ce re ales 29, 513. technologiques du gluten de ble Dervilly-Pinel, G., Thibault, J.-F., Saulnier, L., 2001. Experimental evidence for a semi-exible conformation for arabinoxylans. Carbohydrate Research 330, 365372. lu, F., Aytac ndem-Makasciog Dik, T., Yo , C.H., Kincal, N.S., 2002. Wet separation of wheat ours into starch and gluten fractions: the combined effects of water to our ratio-dough maturation time and the effects of our aging and ascorbic acid addition. Journal of the Science of Food and Agriculture 82, 405413. Eliasson, A.-C., Larsson, K., 1993. Cereals in Bread making. A Molecular Colloidal Approach. Marcel Dekker, New York. Endo, S., Negishi, Y., Shiiba, K., 1989. New modied gluten product and bread improver composition. US Patent, 4,871,577. FAO, 2002. Current production and crop prospects. Food outlook No. 2. http://www.fao.org. ndig, W., Neukom, H., 1963. Ferulic acid as a component of Fausch, H., Ku a glycoprotein from wheat our. Nature 199, 287. Fellers, D.A., 1973. Fractionation of wheat into major components. Symposium Proceedings. AACC, Missouri. Fincher, G.B., Sawyer, W.H., Stone, B.A., 1974. Chemical and physical properties of an arabinogalactan-peptide from wheat endosperm. Biochemical Journal 139, 535545. Frederix, S.A., Courtin, C.M., Delcour, J.A., 2002. The inuence of arabinoxylans on gluten coagulation during a batter process. Poster al, Que bec, Canada. Presentation, AACC Annual Meeting, Montre Frederix, S.A., Courtin, C.M., Delcour, J.A., 2003. Impact of xylanases with different substrate selectivity on glutenstarch separation of wheat our. Journal of Agricultural and Food Chemistry 51, 73387345. Frederix, S.A., Courtin, C.M., Delcour, J.A., 2004a. Inuence of process parameters on yield and composition of gluten fractions obtained in a laboratory scale dough batter procedure. Journal of Cereal Science 39, 2936. Frederix, S.A., Courtin, C.M., Delcour, J.A., 2004b. Substrate selectivity and inhibitor sensitivity affect xylanase functionality in wheat our glutenstarch separation. Journal of Cereal Science, 40, 4149. Fu, B.X., Sapirstein, H.D., 1996. Procedure for isolating monomeric proteins and polymeric glutenin of wheat our. Cereal Chemistry 73, 143152. Goesaert, H., Brijs, K., Veraverbeke, W.S., Courtin, C.M., Gebruers, K., Delcour, J.A., 2004. Wheat our constituents: how they impact bread quality, and how to impact their functionality. Trends in Food Science and Technology, in press. Gruppen, H., Kormelink, F.J.M., Voragen, A.G.J., 1993. Enzymic degradation of water-unextractable cell wall material and arabinoxylans from wheat our. Journal of Cereal Science 18, 129143. Hamer, R.J., Weegels, P.L., Marseille, J.P., Kelfkens, M., 1989. A study of the factors affecting the separation of wheat our into starch and gluten, In: Pomeranz, Y. (Ed.). Wheat is Unique: Structure, Composition, Processing, End-use Properties, and Products. AACC, St Paul, MN, pp. 467477. Hoseney, R.C., 1994. Principles of Cereal Science and Technology, second ed. Association of Cereal Chemists, St Paul, MN. Hlynka, I., 1962. Inuence of temperature, speed of mixing, and salt on some rheological properties of dough in the farinograph. Cereal Chemistry 39, 286303. Izydorczyk, M.S., Biliaderis, C.G., 1995. Cereal arabinoxylans: advances in structure and physicochemical properties. Carbohydrate Polymers 28, 3348.

235

Jelaca, S.L., Hlynca, I., 1971. Water binding capacity of wheat our crude pentosans and their relation to mixing characteristics of dough. Cereal Chemistry 48, 211222. Johnston, P.H., Fellers, D.A., 1971. Process for proteinstarch separation in wheat our, 2. Experiments with a continuous decanter-type centrifuge. Journal of Food Science 36, 649652. Karlsson, R., Olered, R., Eliasson, A.-C., 1983. Changes in starch granule size distribution and starch gelatinisation properties during develop rke 35, ment and maturation of wheat, barley and rye. Starch-Sta 335340. Kelfkens, M., Hamer, R.J., 1991a. Agronomic factors related to the quality rke 43, of wheat for the starch industry. Part I: sprout damage. Starch-Sta 340343. Kelfkens, M., Hamer, R.J., 1991b. Agronomic factors related to the quality of wheat for the starch industry. Part II: nitrogen fertilisation and overall rke 43, 344347. conclusions. Starch-Sta Kema, I.P., Helmens, H.J., Steeneken, P.A.M., 1996. Wet-milling/ammonia rke process for the manufacture of wheat starch and gluten. Starch-Sta 48, 279285. hrmann, C., 1989. Process for the industrial production of Kempf, W., Ro wheat starch form whole wheat. In: Pomeranz, Y. (Ed.). Wheat is Unique: Structure, Composition, Processing, End-use Properties, and Products. AACC, St Paul, MN, pp. 521540. Kerkkonen, H.K., Laine, K.M.J., Alanen, M.A., Renner, H.V., 1976. Method of separation gluten from wheat our. US Patent, 3,951,938. Kerr, R.W., 1950. The manufacture of corn starch, In: Kerr, R.W. (Ed.), Chemistry and Industry of Starch, second ed. Academic Press, New York, pp. 2961. Kim, H.R., Bushuk, W., 1995. Salt sensitivity of acetic acid-extractable proteins of wheat our. Journal of Cereal Science 21, 241250. Kim, S.K., DAppolonia, B.L., 1977. Bread staling studies. III. Effect of pentosans on dough, bread and bread staling rate. Cereal Chemistry 54, 225229. Knight, J.W., Olson, R.M., 1984. Wheat starch: production, modication, and uses. In: Whistler, R.L., Paschall, E.F. (Eds.). Starch: Chemistry and Technology Industrial Aspects, vol. II. Academic Press, New York, pp. 491506. Larsson, H., 1997. Wheat our dough. Rheological and structural aspects. Doctoral Dissertation. Lund University, Sweden. Larsson, H., 2002. Effect of pH and sodium chloride on wheat our dough properties: ultracentrifugation and rheological measurements. Cereal Chemistry 79, 544545. Larsson, H., Eliasson, A.C., 1996a. Phase separation of wheat our dough studied by ultracentrifugation and stress relaxation. I. Inuence of water content. Cereal Chemistry 73, 1824. Larsson, H., Eliasson, A.C., 1996b. Phase separation of wheat our dough studied by ultracentrifugation and stress relaxation. II. Inuence of mixing time, ascorbic acid, and lipids. Cereal Chemistry 73, 2531. Larsson, H., Eliasson, A.C., 1997. Inuence of the starch granule surface on the rheological behaviour of wheat our dough. Journal of Texture Studies 28, 487501. Lineback, D.R., Rasper, V.F., 1988. Wheat carbohydrates. In: Pomeranz, Y. (Ed.). third ed Wheat, Chemistry and Technology, vol. 1. AACC, St Paul, MN, pp. 277372. LMC International Ltd., 2002. Evaluation of the community policy for starch and starch products. http://www.ienica.net usefulreports/starch.pdf. Mecham, D.K., Weinstein, N.E., 1952. Lipid binding in doughs. Effects of dough ingredients. Cereal Chemistry 29, 448455. Meuser, F., Althoff, F., Huster, H., 1989. Developments in the extraction of starch and gluten from wheat our and wheat kernels. In: Pomeranz, Y. (Ed.). Wheat is Unique: Structure, Composition, Processing, End-use Properties, and Products. AACC, St Paul, MN, pp. 479499. Moon, M.H., Giddings, J.C., 1993. Rapid separation and measurement of particle size distribution of starch granules by sedimentation/ steric eld-ow fractionation. Journal of Food Science 58, 11661171.

236

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237 man, P., 1997. A study of the polysaccharide Saulnier, L., Andersson, R., A components in gluten. Journal of Cereal Science 25, 121127. Sayaslan, A., 2004. Wet-milling of wheat our: industrial processes and small-scale test methods. Lebensmittel-Wissenschaft undTechnologie 37, 499515. Schurgers, B., Veraverbeke, W.S., Dornez, E., Delcour, J.A., 2004. Critical fractors governing gluten protein agglomeration on a micro-scale. In: Laandra, D., Masa, S., DOvidio, R. (Eds.), The Gluten Proteins. Proceedings of the VIIIth Gluten Workshop. The Royal Society of Chemistry, Cambridge,. Uk, pp. 292295. Scott, C.S., 1991. Composition of a steeped starched-containing grain and a cellulase enzyme. US Patent, 5,066,218. Seguchi, M., 1993. Effect of wheat our aging on starch-granule surface proteins. Cereal Chemistry 70, 362364. Seguchi, M., Hayashi, M., Kanenaga, K., Ishihara, C., Noguchi, S., 1998. Springiness of pancake and its relation to binding of prime starch to tailings in stored wheat our. Cereal Chemistry 75, 3742. Shewry, P.R., 2003. Wheat gluten proteins. In: Shewry, P.R., Lookhart, G.L. (Eds.). Wheat Gluten Protein Analysis. AACC, St Paul, MN, pp. 117. Sindic, M., Chevalier, O., Duculot, J., Foucart, M., Derouanne, C., 1993. tale et phytotechFractionnement du froment dhiver: inuences varie re ales 2, 37. nique. Industries des Ce Sriroth, K., Walapatit, S., Chollakup, R., Chotineeranat, S., Piyachomkwan, K., Oates, C.G., 1999. An improved dewatering performance in cassava starch process by a pressure lter. Starch rke 51, 383388. Sta Vaara, T., Vaara, M., Simell, M., Lehmussaari, A., Caransa, A., 1989. A process for steeping cereals with a new enzyme preparation. EP 0 321 004 A1. Van den Bulck, K., Loosveld, A.-M.A., Courtin, C.C., Proost, P., Van Damme, J., Robben, J., Mort, A., Delcour, J.A., 2002. Amino acid sequence of wheat our arabinogalactan-peptide, identical to part of grain softness protein GSP-1, leads to improved structural model. Cereal Chemistry 79, 329331. Veal, E.F., 1954. The effect of mixing at selected speeds and temperatures on the physical properties and baking behaviour of doughs made from a Northwest wheat our. Master Thesis. Oregon State University (see http://food.oregonstate.edu/ref/bake/veal). Veraverbeke, W.S., Delcour, J.A., 2002. Wheat protein composition and properties of wheat glutenin in relation to breadmaking functionality. Critical Reviews in Food Science and Nutrition 42, 179208. Verberne, P., Zwitserloot, W., 1978. A new hydrocyclone process for the separation of starch and gluten from wheat our. Starch-Starke 30, 337 338. Wadhawan, C.K., 1988. Fundamental studies on vitality of gluten for breadmaking. PhD Dissertation, Winnipeg, Manitoba. Wang, M., Hamer, R.J., Van Vliet, T., Oudgenoeg, G., 2002. Interaction of water extractable pentosans with gluten protein: effect on dough properties and gluten quality. Journal of Cereal Science 36, 2537. Wang, M., Hamer, R.J., Van Vliet, T., Gruppen, H., Marseille, H., Weegels, P.L., 2003. Effect of water unextractable solids on gluten formation and properties: mechanistic considerations. Journal of Cereal Science 37, 5564. Watson, S.A., 1964. Preparations of starch and starch fractions. In: Whistler, R.L. (Ed.). Whole Starch. Corn Starch Methods in Carbohydrate Chemistry, vol. IV. Academic Press, New York, pp. 35. Watson, S.A., 1967. Manufacture of corn and milo starches. In: Whistler, R.L., Paschall, E.F. (Eds.). Starch: Chemistry and Technology Industrial Aspects, vol. II. Academic Press, New York, pp. 151. Watson, S.A., Williams, C.B., Wakely, R.D., 1951. Laboratory steeping procedures used in a wet milling research program. Cereal Chemistry 28, 105118. Weegels, P.L., Marseille, J.P., de Jager, A.M., Hamer, R.J., 1990. Structurefunction relationships of gluten proteins. In: Bushuk, W., Tkachuk, R. (Eds.). Gluten Proteins. American Association of Cereal Chemists, St Paul, MN, pp. 98111.

Olcott, H.S., Mecham, D.K., 1947. Characterization of wheat gluten. I. Proteinlipid complex formation during doughing of ours. Lipoprotein nature of the glutenin fraction. Cereal Chemistry 24, 407414. Olsen, H. 2002. A method for the separation of wheat our using a transglutaminase enzyme. WO 02/15713 A1. Olsen, H., Nielsen, B., 2002. Starch gluten separation process. WO 02/00911 A1. Osborne, T.B., 1924. The Vegetable Proteins. Longmans & Green, London. Perlin, A.S., 1951a. Isolation and composition of the soluble pentosans of wheat our. Cereal Chemistry 28, 370381. Perlin, A.S., 1951b. Structure of the soluble pentosans of wheat ours. Cereal Chemistry 28, 282393. Petrofsky, K.E., Hoseney, R.C., 1995. Rheological properties of dough made with starch and gluten from several cereal sources. Cereal Chemistry 72, 5358. Picout, D.R., Ross-Murphy, S.B., 2002. On the chain exibility of arabinoxylans and other b-(1/4) polysaccharides. Carbohydrate Research 337, 17811784. Prieto, J.A., Kelfkens, M., Weegels, P.L., Hamer, R.J., 1992. Variations in the gliadin pattern of our and isolated gluten on nitrogen applicationimplications for baking potential and rheological properties. Zeitschrift fur Lebensmittel-Untersuchung und -Forschung 194, 337343. Rao, G.V., 1979. Wet wheat milling. Cereal Foods World 24, 334335. Rao, G.V., Gerrish, O.B., 1975. Process for separation of whole wheat kernel components to isolate the gluten employing water. US Patent, 3,891,613. Rao, G.V., Shoup, F.K., 1978. Method for fractionating the whole wheat kernel by centrifugal means. US Patent, 4,125,528. Rao, G.V., Henry, W.E., Hammond, D.L., 1976. Fractionation of the whole wheat kernel by pin milling. US Patent, 3, 979375. Redgwell, R.J., de Michieli, J.-H., Fischer, M., Reymond, S., Nicolas, P., Sievert, D., 2001. Xylanase induced changes to water- and alkaliextractable arabinoxylans in wheat our: their role in lowering batter viscosity. Journal of Cereal Science 33, 8396. Redman, D.G., 1971. Softening of gluten by wheat proteases. Journal of the Science of Food and Agriculture 22, 75. Robertson, G.H., Cao, T., 1998a. Substitution of concentrated ethanol for water in the laboratory washing fractionation of protein and starch from hydrated wheat our. Cereal Chemistry 75, 508513. Robertson, G.H., Cao, T.K. 1998b. Methods for separation of wheat our into protein and starch fractions. US Patent, 5,851,301. Robertson, G.H., Cao, T.K., Ong, I., 1999. Wheat gluten swelling and partial solubility with potential impact on starch-from-gluten separation by ethanol washing. Cereal Chemistry 76, 843845. Rodgers, N.E., Gidlow, R.G., 1974. Hydroprocessing of wheat. US Patent, 3,851,085. Roels, S.P., 1997. Factors governing wheat and wheat gluten functionality in breadmaking and gluten/starch separation. PhD Dissertation. Katholieke Universiteit Leuven, Belgium. Roels, S.P., Cleemput, G., Vandewalle, X., Nys, L., Delcour, J.A., 1993. Bread volume potential of variable-quality ours with constant protein level is determined by factors governing the mixing time and baking absorption levels. Cereal Chemistry 70, 318323. Roels, S.P., Grobet, P.J., Delcour, J.A., 1998a. Distribution of carbohydrates in gluten fractions isolated from European wheats (Triticum aestivum L.) in a batter system. Journal of Agricultural and Food Chemistry 46, 13341343. Roels, S.P., Sindic, M., Deroanne, C., Delcour, J.A., 1998b. Protein composition and agglomeration tendency of gluten isolated from European wheats (Triticum aestivum L.) in a batter system. Journal of Agricultural and Food Chemistry 46, 13441349. Rouau, X., El-Hayek, M.-L., Moreau, D., 1994. Effect of an enzyme preparation containing pentosanases on the bread-making quality of ours in relation to changes in pentosan properties. Journal of Cereal Science 19, 259272.

A. Van Der Borght et al. / Journal of Cereal Science 41 (2005) 221237 Weegels, P.L., Marseille, J.P., Hamer, R.J., 1988. Small scale separation of rke 40, 342346. wheat our in starch and gluten. Starch-Sta Weegels, P.L., Marseille, J.P., Hamer, R.J., 1992. Enzymes as a processing rke aid in the separation of wheat our into starch and gluten. Starch-Sta 44, 4448. Wellner, N., Bianchini, D., Mills, E.N.C., Belton, P.S., 2003. Effect of selected Hofmeister anions on the secondary structure and dynamics of wheat prolamins in gluten. Cereal Chemistry 80, 596600. Wootton, M., Mahdar, D., 1993. Properties of starches from Australian rke 45, wheats. Part 1: separation of starch and gluten. Starch-Sta 255258.

237

lu, F., Dik, T., Kincal, N.S., 2002. Separation of bread ndem-Makasciog Yo wheat ours into starch and gluten fractions: effect of water temperature alone or in combination with water to our ratio. Journal of the Science of Food and Agriculture 82, 414420. Yuan, J., Chung, D.S., Seib, P.A., Wang, Y., 1998. Effect of steeping conditions on wet-milling characteristics of hard red winder wheat. Cereal Chemistry 75, 145148. Zawistowska, U., Bekes, F., Bushuk, W., 1985. Gluten proteins with high afnity to our lipids. Cereal Chemistry 62, 284289. Zhuge, Q., Persaud, J.N., Posner, E.S., Deyoe, C.W., Seib, P.A., Chung, D.S., 1991. Isolation of gluten and starch from ground, pearled wheat compared to isolation from our. Cereal Chemistry 68, 336399.