A bench-scale high-shear wet-milling test for wheat flour

(1)

A bench-scale high-shear wet-milling test for wheat ﬂour

Abdulvahit Sayaslan

a,⇑

, Paul A. Seib

b

, Okkyung Kim Chung

c

a

Department of Food Engineering, Karamanog˘lu Mehmetbey University, Karaman, Turkey b

Department of Grain Science and Industry, Kansas State University, Manhattan, KS, USA c

USDA/ARS, Grain Marketing and Production Research Center, Manhattan, KS, USA

a r t i c l e

i n f o

Article history:

Received 19 October 2011

Received in revised form 6 February 2012 Accepted 11 February 2012

Available online 20 February 2012 Keywords:

Wheat ﬂour Wet-milling quality High-shear test

a b s t r a c t

Bench-scale wet-milling tests for wheat flour are available for the traditional processes of Martin and Batter, but tests for the high-shear processes of Alfa-Laval/Raisio, Hydrocyclone, and High-Pressure Dis-integration are few in number. In this study, critical processing parameters of a high-shear wet-milling process, namely high-shear mixing, gluten-aging, and gluten-washing steps, were investigated using response surface methodology, and those parameters led to a bench-scale wet-milling test starting with a ‘‘highly sheared flour–water dispersion’’ (HS-FWD). Optimum conditions for the test were: a water– flour ratio (db) of 1.7, water temperature of 35 °C, and homogenizer speed of 6000 rpm for 2.0 min in the high-shear mixing step, a temperature of 40 °C for 20 min in the aging step, and gluten-washing of 2.0 min in the Glutomatic system. The HS-FWD wet-milling test with high level of repeatabil-ity was capable of discriminating wet-milling qualities of several hard, soft, and coarsely ground wheat flours.

1. Introduction

Wheat has become a significant raw material for the co-production of wheat starch and vital wheat gluten via wet-milling, especially in Europe during the last several decades (Lindhauer, 1997; Witt, 1997; Sayaslan, 2004). Wheat starch and vital gluten are used in numerous food, pet-food, and nonfood applications (Maningat et al., 1994; Bergthaller, 1997; Maningat and Seib, 1997). Wet-separation and purification of gluten proteins and starch granules from wheat flour are based principally on differ-ences in their aggregation, particle sizes, and sedimentation rates (Sayaslan, 2004; Van Der Borght et al., 2005).

Five commercial wet-milling processes, namely the Martin (dough-washing), the Batter, the Alfa-Laval/Raisio, the Hydrocy-clone, and the High-Pressure Disintegration (HD) processes, have been employed to produce wheat starch and vital gluten, and those wet-processes start with ﬂour rather than wheat kernels (Seib, 1994; Maningat and Bassi 1999; Sayaslan, 2004). The Martin and Batter processes are considered old traditional processes, whereas the other three are relatively new (Meuser et al., 1989; Sayaslan, 2004). In the traditional processes, separation of ﬂour components starts with a stiff dough (Martin process) or a slack batter (Batter process), then proceeds with kneading of the dough and/or screening to separate starch granules from gluten ‘‘curds’’ (Roels et al., 1998a,b; Sayaslan, 2004; Van Der Borght et al., 2005). In

the Martin process, the gluten proteins are strongly aggregated during dough (40% mc) mixing to form sheets of gluten strands that enmesh other dough components. The gluten proteins are also aggregated in the mixed Batter (65% mc) due to the warm water (50 °C) added in the mixing step (Anderson, 1967, 1974; Fellers, 1973; Knight and Olson, 1984; Seib, 1994). However, in the modern-day processes, separation starts with a ‘‘highly sheared ﬂour–water dispersion’’ (HS-FWD) or a ‘‘moderately sheared dough–water dispersion’’ (MS-DWD), wherein a reduced degree of gluten development is affected. The dispersion proceeds to cen-trifugation, where starch granules sediment through a matrix of relatively thin gluten strands (Sayaslan, 2004).

Of the modern-day processes, the Alfa-Laval/Raisio and the HD processes both begin with a HS-FWD, whereas the Hydrocyclone process begins with a MS-DWD. The HS-FWD and MS-DWD apparently possess thin and relatively short threads of gluten to-gether with many free starch granules that are dispersed in water (Sayaslan, 2004; Sayaslan et al., 2010). The mechanical and fluid shear forces used in those processes accomplish two important tasks. First, the indigenous association between starch granules and protein matrix in a flour particle is replaced by an association with water. Second, mixing causes hydrated gluten molecules to collide and aggregate into threads that align in the shear field ( Ver-berne and Zwitserloot, 1978; Zwitserloot, 1989; Meuser, 1994; Witt, 1997; Bergthaller et al., 1998). As opposed to the traditional processes, in which gluten proteins are aggregated into curd-sizes during mixing, the high-shear forces used in the modern-day processes restrict the size of gluten aggregates to thin

⇑ Corresponding author. Tel.: +90 544 212 8856; fax: +90 338 226 2023. E-mail address:sayaslan@kmu.edu.tr(A. Sayaslan).

Contents lists available atSciVerse ScienceDirect

Journal of Food Engineering

(2)

thread-shaped particles having lengths in the micro to millimeter range (Robertson and Cao, 1998a,b; Sayaslan et al., 2010). Due to the advantages of reduced water usage and shortened residence time during processing, the modern-day processes are preferred over the traditional ones by the wheat wet-milling industry (Sayaslan, 2004; Van Der Borght et al., 2005).

Several bench-scale and/or pilot-scale tests are available to as-sess wet-milling quality of wheat flour by the Martin (Schafer, 1975; Godon et al., 1983) and the Batter processes (Weegels et al., 1988; Hamer et al., 1989; Wang et al., 2002, 2003a,b; Wang, 2003). However, test methods for the modern-day shear processes are few. Bergthaller et al. (1998)andLindhauer and Bergthaller (2002) reported a ‘‘mixer test’’ to simulate the high-shear pro-cesses starting with 100 g of flour. However, the sheared flour– water mixture was treated in the ‘‘mixer test’’ as practiced in the Batter process (dilute the sheared mixture with warm water and allow gluten aggregation in the presence of all flour components) as opposed to that practiced in the modern-day processes (centri-fuge and separate a protein-rich fraction from which gluten is aggregated).Bergthaller et al. (1998)also devised a pilot-scale test (20 kg of flour) starting with a special mixing unit that was re-ported to be of a screw design and capable of continuously shear mixing the flour with water to give a flowable dispersion. The resulting HS-FWD was fractionated using a 2-phase decanter cen-trifuge of pilot-scale size.

Given the release of numerous wheat varieties each year, plus the breeding of waxy and high-amylose specialty wheats in recent years, and the preference of high-shear processes by the ing industry, it is important that bench-scale high-shear wet-mill-ing tests be available. The aims of this research were to investigate the critical parameters in a high-shear wet-milling test for wheat flour, and design a bench-scale (<100 g of flour) HS-FWD wet-mill-ing test to differentiate the wet-millwet-mill-ing qualities of flours.

2. Materials and methods

2.1. Materials

A regional baking standard (RBS) flour for pan bread with 11.8% protein (14% mb), obtained from the USDA/ARS Grain Marketing and Production Research Center (GMPRC), currently the Center for Grain and Animal Health Research, Manhattan, KS, was used to de-vise the HS-FWD wet-milling test. Four hard red winter wheats, namely Karl 92, Jagger, 2137, and TAM 110, which, respectively had extra strong, strong, medium, and weak dough-mixing proper-ties (Fig. 1), were obtained from the Agricultural Experiment Station of Kansas State University, Manhattan, KS. A hard red winter wheat (Chisholm) flour with extra long mixing time (Fig. 1) was provided by the USDA/ARS GMPRC, Manhattan, KS. A soft red winter wheat (AT 91W) flour (Fig. 1) was acquired from the USDA/ARS Soft Wheat Quality Laboratory, Wooster, OH. A coarsely ground commercial flour milled through a short-flow dry-milling process (Fig. 1) was from Midwest Grain Products, Inc., currently MGP Ingredients, Inc., Atchison, KS. All wheat samples were tempered at 15.5% mois-ture for 16 h and roller-milled to straight-grade flours on a Buhler experimental mill (M202, Buhler Co., Uzwil, Switzerland).

2.2. General methods

Moisture contents were determined by AACC Method 44-15A (AACC, 2000). Dry-solid contents of the water-solubles fractions were determined by oven-drying of an aliquot (20 mL) at 105 °C for 3 h after water was evaporated from the sample at 80–90 °C overnight. Dry-solid contents of the ﬁber fractions were deter-mined by oven drying of the whole fractions (20 g) at 105 °C

for 3 h. Protein (Nx5.7) contents of the RBS flour and of wet-milling fractions were determined by the FP Protein/Nitrogen Analyzer (Leco Corp., St. Joseph, MI). Total starch and damaged starch con-tents of the RBS flour and starch fractions were assayed by AACC Methods 76-13 and 76-31, respectively, using assay kits from Megazyme International Ireland Ltd., Wicklow, Ireland. Insoluble polymeric protein contents of flours were determined by the meth-od ofBean et al. (1998). Soluble proteins and water-soluble solids in a flour at 35 °C were determined as follows: flour (10.0 g, db) and water (100 mL at 35 °C) were combined in a 250-mL centrifuge bottle, shaken at a medium speed for 10 min in a reciprocal water bath (Precision Instruments, Winchester, VA) maintained at 35 °C. The slurry was centrifuged at 2500g for 15 min, and an aliquot of the supernatant was assayed for protein and dry-solids. Mixing properties of flours were determined by AACC Method 54-40A using a 10-g mixograph (National Manufacturing Co., Lincoln, NE). The mixograms were interpreted as described byFinney and Shogren (1972). Scanning electron micrographs (SEM) of freeze-dried samples of a flour–water dough, a moderately sheared dough–water dispersion (MS-DWD), and a highly sheared flour–water dispersion (HS-FWD) were acquired with a Hitachi S-3500N (Hitachi Science Systems, Ltd., Tokyo, Japan) scanning electron microscope operated at 5 kV.

Fig. 1. Mixograms at optimum water absorption levels of ﬂours with different dough-mixing properties (protein contents are based on 14% moisture content).

(3)

2.3. Development of a bench-scale HS-FWD wet-milling test

2.3.1. Optimization of the ‘‘high-shear mixing’’ step in the HS-FWD wet-milling test

A three-way (3 independent variables with 3 levels each) re-sponse surface methodology (RSM) was used, where the RSM de-sign (Table 1) specified a total of 20 randomized experimental runs (Walker and Parkhurst, 1984). Three factors, namely water/ flour (W/F) ratio, water temperature, and homogenizer speed (shear), were selected as the independent variables. Each variable consisted of three levels: W/F ratios (db) of 1.4, 1.7, and 2.0, water temperatures of 20, 27, and 34 °C, and homogenizer speeds of 2500, 5000, and 7500 rpm. The homogenizer was a Model Pro 350PC from Pro Scientific, Monroe, CT, and the shearing (homoge-nization) time was fixed at 2.0 min based on initial trials. Flour was at room temperature. Measured responses were the temperature of the HS-FWD after homogenization, recovery of starch in the A-starch fraction, and recovery of protein in the gluten fraction. The recovery of starch (%) was calculated as follows: 100 (mass of starch fraction) (weight fraction of starch in starch fraction)/(mass of flour) (weight fraction of starch in flour). The recovery of protein was calculated in a similar fashion. The collected data were ana-lyzed by RSM-PlusÓ Enhanced Response Surface Methodology Analyzer and General Equation Modeler (AEW Consulting, Lincoln, NE) software to fit to a second-order (quadratic) multiple regres-sion model for each response Y ¼ A þ BXð 1þ CX2þ DX3þ EX1X2þ

FX1X3þ GX2X3þ HX21þ IX 2

2þ JX

2

3Þ. The RSM-PlusÓ software

pro-vided coefﬁcient of determination (R2_{) and standard error of}

esti-mate for models that reﬂect the goodness of the ﬁt of the data to the models.

Details of the wet-milling test, outlined inFig. 2, are as follows: ﬂour (65.0 g, db) and water (amount and temperature as per

experimental design) were measured. Approximately one-half of the water was placed in a 250-mL clear wide-mouth centrifuge bottle (Cat. No. 05-432A, Fisher Scientific, Chicago, IL) and then the flour followed by the remaining water were added. In that manner, formation of lumps of flour on the bottom of the centri-fuge bottle was prevented. The flour and water were gently blended with a spatula for 5–10 s and the slurry was subjected to high-shear mixing with a homogenizer (Model Pro 350PC) equipped with a 37-mm diameter rotor–stator generator at speeds of 2500, 5000, or 7500 rpm (as per experimental design) for 2.0 min. During the homogenization stage, the centrifuge bottle was manually moved vertically at about 30 cycles per minute. The temperature of the HS-FWD was recorded upon completion of the shearing. The rotor–stator generator was rinsed by a water stream (20 mL, 25 °C) from a wash bottle and the washings were added to the bottle containing the dispersion. After the solids con-centration of the HS-FWD was adjusted to 27% solids by adding 55 mL of water at 25 °C, the centrifuge bottle was capped and the contents mixed by inverting the bottle manually 10 times. The HS-FWD was then centrifuged at 2500g for 15 min using a swinging-bucket rotor (Model JS 7.5) centrifuge (Model J2-21, Beckman Coulter, Inc., Palo Alto, CA). Three major phases became visible in the centrifuge bottle: a water-solubles phase (top), a pro-tein-rich phase (middle), and a starch phase (bottom). The top water-solubles phase was decanted and saved to combine with other water-solubles generated during the process. It was noted that the protein-rich phase was topped by a thin layer of sedi-mented gel, which was probably comprised of small and damaged starch granules along with insoluble-pentosans. The thin gel phase was not clearly visible but was readily differentiated manually by its soft texture as opposed to the strong elastic nature of the under-lying gluten proteins. At the bottom of the protein-rich phase and

Table 1

Response surface methodology (RSM) models of the high-shear mixing step for the temperature of the sheared flour–water dispersion, the recovery of starch in the A-starch fraction and the recovery of protein in the gluten fraction in the highly sheared flour–water dispersion (HS-FWD) wet-milling test on the RBS wheat flour.

Run no. RSM experimental variables Experimental and predicted responses Water/ﬂour ratioa

Homogen. speed (rpm) Water Temp. (°C) Temperature of HS-FWD (°C)

Starch recovery in A-starch fractionc_{(% of ﬂour starch)}

Protein recovery in gluten fractione_{(% of ﬂour protein)}

(a) (b) (c) Exp. Predictedb

Exp. Predictedd Exp. Predictedf 1 1.4 7500 20 36.0 36.0 76.4 77.2 65.5 66.9 2 1.7 2500 20 27.5 27.5 68.8 70.1 66.7 66.7 3 2.0 5000 20 28.5 29.5 80.3 78.6 63.8 65.2 4 1.7 5000 27 32.7 33.1 76.5 79.6 65.2 66.3 5 1.7 5000 27 33.5 33.1 79.2 79.6 66.1 66.3 6 1.7 5000 27 34.0 33.1 81.1 79.6 67.5 66.3 7 2.0 7500 27 39.8 39.1 83.9 84.1 66.1 65.3 8 1.4 5000 34 38.0 38.3 73.1 72.6 69.5 69.6 9 1.7 7500 34 40.5 39.7 80.8 80.7 65.9 67.2 10 2.0 2500 34 34.0 34.4 60.4 56.4 69.4 68.4 11 1.4 7500 20 35.9 36.0 78.0 77.2 68.3 66.9 12 1.7 2500 20 27.5 27.5 71.4 70.1 66.6 66.7 13 2.0 5000 20 30.5 29.5 76.9 78.6 66.6 65.2 14 1.7 5000 27 32.5 33.1 80.0 79.6 66.1 66.3 15 1.7 5000 27 32.5 33.1 80.9 79.6 66.1 66.3 16 1.7 5000 27 33.2 33.1 80.0 79.6 66.7 66.3 17 2.0 7500 27 38.4 39.1 84.2 84.1 64.4 65.3 18 1.4 5000 34 38.5 38.3 72.1 72.6 69.7 69.6 19 1.7 7500 34 38.8 39.7 80.5 80.7 68.4 67.2 20 2.0 2500 34 34.8 34.4 52.3 56.4 67.3 68.4

a _{W/F ratio is based on dry weight of ﬂour. Flour was the regional baking standard (RBS) ﬂour.}

b _{RSM regression equation for temperature (°C): 31.46 + 14.86a 0.0029b + 3.957c + 0.00070ab 0.93ac 0.00011bc + 0.85a}2_{+ 0.0000006b}2

0.025c2_{(Coefﬁcient of} determination R2

: 0.98; Standard error of estimate: 0.82). c

Starch recovery (%): 100 (mass of starch fraction) (mass fraction of starch in starch fraction)/(mass of ﬂour) (mass fraction of starch in ﬂour). d

RSM regression equation for starch recovery (%): 30.88 + 46.89a + 0.0029b + 4.43c + 0.00017ab 0.56ac + 0.00019bc 9.33a2

0.0000005b2 0.089c2

(Coefﬁcient of determination R2

: 0.95; Standard error of estimate: 2.41). e

Protein recovery (%): 100 (mass of gluten fraction) (mass fraction of protein in gluten fraction)/(mass of ﬂour) (mass fraction of protein in ﬂour). f

RSM regression equation for protein recovery (%): 61.90 – 9.90a 0.00021b + 1.14c 0.00089ab - 0.39ac 0.000009bc + 6.18a2

+ 0.0000002b2

0.0052c2

(Coefﬁcient of determination R2_{: 0.65; Standard error of estimate: 1.33).}

(4)

atop the starch phase was a thin layer of ﬁber. The protein-rich phase together with its thin gel atop was carefully scraped off with a spatula, then slurried with 100 mL of water at 25 °C. After aging (maturing) at 40 °C for 20 min, the mixture was added atop a set of three 8-inch vibrating sieves with openings of 425, 250, and 63

l

m (US Standard Sieves No. 40, 60, and 230) (Model Fritsch A-3, Gilson Company, Inc., Lewis Center, OH). The overs on the 425- and

250-l

m sieves were combined and divided in two equal portions. The two portions were added each to a Glutomatic washing chamber ﬁtted with a 88-

l

m opening screen. The Glutomatic instrument

(Model 2200, Perten Instruments North America, Inc., Reno, NV) had been prerun for 20 s before mounting the chambers because the dough-mixing step normally used to isolate gluten from ﬂour was omitted. The instrument was put on hold at the starting time of the gluten-washing step and the chambers containing the gluten concentrate phase were mounted. The gluten isolation and puriﬁ-cation step was then initiated and continued at a preset water stream of 50–60 mL/min for 2.0 min. The wet gluten portions re-tained inside the two chambers were combined together, then added to two other gluten streams, one from the gluten recovered

(5)

on top of two sieves (425 and 250

l

m) after the vibratory wet-siev-ing of the gluten washwet-siev-ings, and the other from the wet-sievwet-siev-ing of the slurry of the starch phase. The combined wet glutens were placed in a thick-walled centrifuge bottle, expanded to several vol-umes under low pressure, and ﬁnally dried on a freeze-dryer (Flexi-Dry/MP, TMS Systems, Inc., Stone Ridge, NY). The dried glu-ten fraction was ground on a Thomas Wiley Mill (Cat. No. 08-338, Fisher Scientiﬁc) to pass through a 420-

l

m opening sieve. The dried gluten fraction was analyzed for moisture and protein.

The starch phase was mixed uniformly in 75 mL of water by vig-orous hand-shaking. The slurry was subjected to wet-sieving at full amplitude (3.0 mm) on the vibrating sieve-shaker. The 425- and 250-

l

m screens retained mostly small gluten particles, which were combined with the wet gluten generated by the Glutomatic washing as described above. The 63-

l

m screen retained mainly the fine fiber and insoluble pentosans, which were called fiber frac-tion. After adding the slurry to the top screen, tap water was sprayed on the top screen at a rate of 700 mL/min during the first 30 s of wet-sieving. The rinse water was turned off and sieving con-tinued another 90 s. The throughs were collected and transferred to two 250-mL bottles and centrifuged (2500g for 10 min). The supernatants in the bottles were decanted and added to the water-solubles fraction. The sediments in the bottles were slurried in water (75 mL each) and then combined in one bottle. After cen-trifugation, the supernatant was removed and combined with the water-solubles fraction. The top thin layer (1 mm thick) of the sedimented solids, often called tailings, was carefully removed with a spatula and combined with the throughs (washings) gener-ated by the Glutomatic washing of the gluten. The sedimented starch was dried in a forced-air convection oven (35–40 °C) and ground on a laboratory mill (Model A-10, Tekmar–Dohrmann, Ma-son, OH) for 15 s to produce the A-starch fraction, which was as-sayed for protein and moisture. The starch content of a particular starch fraction was calculated to be 100 minus the sum of protein and moisture content.

The throughs from the Glutomatic separation were subjected to vibratory wet-sieving. The overs on the top two screens (425 and 250

l

m) became part of the gluten fraction, whereas those atop the 63-

l

m screen were part of the ﬁber fraction. The throughs from the vibrating sieves were transferred to two 250-mL bottles and centrifuged. The supernatants in the bottles were decanted

and combined with the water-solubles phase, while the sediments in the bottles were slurried in a total of 200 mL of water and the slurry placed in one centrifuge bottle. After centrifugation, the supernatant was combined with the water-solubles, while the sed-iment, called the B-starch fraction, was oven-dried (35–40 °C), ground, and analyzed for moisture and protein. An aliquot (20 mL) of the total water-solubles, which contained all superna-tants and totaled about 1800–2000 mL, was heated in an oven at 80–90 °C overnight to evaporate water and the residue was dried at 105 °C for 3 h. The residual dry-solid was a mixture of water-solubles plus particulate material, but was simply termed the water-solubles. The residue that remained on the 63-

l

m sieve after sieving the fractions was collected and dried at 105 °C for 3 h. That residue was the ﬁber fraction.

2.3.2. Optimization of the ‘‘gluten-aging’’ step in HS-FWD wet-milling test

For optimization of the gluten-aging step in the HS-FWD wet-milling test, a two-way (2 independent variables with 3 levels each) RSM design was followed (Table 2). Two factors, namely temperature of water added to the protein-rich phase in the beaker and aging time, were selected as the independent variables. Each variable consisted of three levels: water temperatures of 25, 35, and 45 °C and aging times of 1, 25, and 45 min. The flour used was the RBS, and the high-shear mixing step on the slurry of flour (65 g, db) and water (100 mL, 35 °C) was always done at 6000 rpm for 2 min. Also, the amount of water added to the protein-rich phase in the beaker was fixed at 100 mL based on initial trials. The recovery of flour protein in the gluten fraction was measured as the response. The collected data were analyzed by the RSM-PlusÓ software to fit to a quadratic multiple regression model

Y ¼ A þ BX1þ CX2þ DX1X2þ EX21þ FX 2 2

.

After centrifugal fractionation of the HS-FWD and decanting the top liquid phase (Fig. 2), the protein-rich phase was scraped off and transferred to a 250-mL beaker with 100 mL of water at 25, 35, or 45 °C as per the experimental design. The beaker was placed in a water bath that was maintained at 25, 35, or 45 °C, and the mixture was rested without agitation for 1, 25, or 45 min as per the exper-imental design. Upon completion of gluten-aging, the contents of the beaker were poured atop the vibrating screens (425, 250, and 63

l

m) and the overs (gluten aggregates) on the 425 and

250-Table 2

Response surface methodology (RSM) models of the aging (maturing) step for the protein-rich phase in the highly sheared ﬂour–water dispersion (HS-FWD) wet-milling test on the RBS wheat ﬂour.a

Run no RSM experimental variables Experimental and predicted responses Aging time (min) Aging temperature (°C) Protein recovery in gluten fractionb

(% of ﬂour protein)

(a) (b) Experimental Predictedc

1 1 25 73.2 73.2 2 1 45 73.2 73.3 3 25 35 75.1 74.9 4 25 35 74.7 74.9 5 25 35 74.9 74.9 6 45 25 72.4 72.5 7 45 45 73.4 73.1 8 1 25 73.1 73.2 9 1 45 73.4 73.3 10 25 35 74.8 74.9 11 25 35 75.0 74.9 12 25 35 74.7 74.9 13 45 25 72.5 72.5 14 45 45 72.7 73.1

a _{The variables were aging temperature and time, and the response was the recovery of flour protein in the gluten fraction.} b _{Protein recovery (%): 100 (mass of gluten fraction) (mass fraction of protein in gluten fraction)/(mass of flour) (mass fraction of} protein in flour).

c

RSM regression equation for protein recovery (%): 66.26 + 0.097a + 0.42b + 0.00051ab 0.0027a2

0.0059b2 (Coefﬁcient of determination R2

(6)

l

m sieves were collected and processed in the Glutomatic as pre-viously described. The wet glutens retained on the two washing chambers of the Glutomatic were combined, freeze-dried, and ground as described above.

2.4. Moderately sheared dough–water dispersion (MS-DWD) wet-milling test

For the MS-DWD test, which resembles the Hydrocyclone pro-cess, the wet-milling procedure of Czuchajowska and Pomeranz (1993, 1995)was modified (Sayaslan, 2002; Sayaslan et al., 2006, 2010). RBS flour (65.0 g, db) was divided into two equal portions and mixed to doughs in two consecutive batches in a 35-g mixog-raph (National Manufacturing Co., Lincoln, NE) at optimum water absorption (63.2%, 14% mb) and optimum mixing time (4.0 min). The doughs were combined and placed in a 300-mL capacity bowl of a Waring blender containing 115 mL of water at 15 °C. After a 30-min resting period, the dough and water were mixed rapidly in the Waring blender for 1.0 min at full speed. The so-called MS-DWD was then transferred to a 300-mL centrifuge bottle and centrifuged at 2500g for 15 min. The vital gluten, A-starch, B-starch, water-solubles, and fiber fractions were isolated and puri-fied as in the HS-FWD wet-milling test. The protein-rich phase col-lected upon centrifugation of the dough-dispersion was washed in the Glutomatic without a maturation step.

2.5. Dough-washing (Martin) wet-milling test

Starting with 65.0 g (db) of ﬂour, a dough was prepared in the same manner as in the MS-DWD method (Sayaslan, 2002; Sayaslan et al., 2010). The dough was covered with a wet cloth and rested at room temperature for 1 h. The rested dough-ball was hand-washed under a stream of water at 25 °C (100 mL/min) with a total water usage of 750 mL. The wet gluten was freeze-dried and ground as described in the HS-FWD test method. The starch milk was wet-sieved on a 63-

l

m sieve to collect the ﬁber, and the throughs were centrifuged (2500g/10 min) in six 250-mL bottles. The superna-tants were decanted and saved, and the sediments were slurried with water (150 mL total) in one bottle and centrifuged. After decanting the supernatant and combining with the previous one, the upper pigmented layer (4–5 mm) was scraped off with a spat-ula and saved, while the sediment was slurried with 150 mL of water and centrifuged. The supernatant was combined with the others, and the combined supernatant was sampled for assay of water-solubles and protein. The upper pigmented layer (1 mm) was scraped off and combined with the previous one to give B-starch, while the sediment was recovered as the A-starch fraction. Both starch fractions were dried and ground as described previ-ously and assayed for protein and moisture contents.

2.6. Statistical analyses

All analyses were carried out at least in two replications and the means were compared by the least signiﬁcant difference (LSD) multiple comparison test at

a

= 0.05 level in one-way analysis of variance (ANOVA), using the Statistical Analysis System Software (SAS Institute, Inc., Cary, NC).

3. Results and discussion

3.1. Properties of RBS ﬂour and microstructures of ﬂour–water dough and sheared dispersions

The RBS ﬂour used to design the HS-FWD test had the average properties of a straight-grade hard wheat ﬂour with 11.8% protein

(14% mb) and 4.7% damaged starch (14% mb). The mixograph data (Fig. 1) and baking properties of the RBS ﬂour, published elsewhere bySayaslan et al. (2010), were as follows: optimum water absorp-tion of 63.2% (14% mb), optimum mixing time of 4.0 min, mixing tolerance of 4 out of 6 (satisfactory), 100-g ﬂour pup-loaf volume of 885 cc and crumb grain score of 4 out of 6 (satisfactory).

As shown inFig. 3, high-shear mixing forces were required to separate the RBS flour components during centrifugation of its flour–water dispersion. When the flour–water mixture was hand-shaken in a centrifuge bottle without high-shear mixing and then centrifuged, only two phases were observed (Fig. 3a). However, when that flour–water mixture was subjected to high-shear forces, as in the HS-FWD test, and then centrifuged, a clear separation of three major phases was observed; the top superna-tant phase, a middle protein-rich phase, and a bottom starch phase (Fig. 3b). It is apparent that the high-shear mixing step of the HS-FWD wet-milling test, wherein the flour and water were mixed at high speed, broke the cohesive forces between starch granules and protein matrix in the flour particles, leading to the release of individual gluten threads and starch granules into the aqueous medium.

SEM photomicrographs of the optimally mixed flour–water dough, MS-DWD, and HS-FWD are shown inFig. 4. In the dough that was mixed at optimum water absorption and mixing time, the gluten was aggregated to form a continuous sheet-like struc-ture that tightly enmeshed the discontinuous starch granules (Fig. 4a). When that dough was rapidly mixed with an excess of water in a blender to form the so-called MS-DWD, the sheet-like network was transformed to thick thread-shaped particles of glu-ten with most starch granules freed of gluglu-ten (Fig. 4b). Those glu-ten threads were observed to be irregularly shaped and sized, and under a centrifugal field, those large-sized particles would sedi-ment together with starch and cause inefficient separation of pro-tein and starch. In the case of the HS-FWD, where the flour and water were highly sheared directly without forming a dough, the gluten network was observed to be an assembly of intercon-nected gluten threads (Fig. 4c). During the centrifugal separation, the fine network of the gluten threads was relatively buoyant and allowed most of the starch granules in the dispersion to pass through and form the almost pure starch phase at the bottom (Fig. 3).

Fig. 3. Centrifugal separation (2500g/15 min) of a non-sheared ﬂour–water slurry and a highly sheared ﬂour–water dispersion (HS-FWD).

(7)

3.2. Modeling of the ‘‘high-shear mixing’’ step in the HS-FWD wet-milling test; recoveries of starch and protein in the A-starch and gluten fractions

Based on the prior work (Zwitserloot, 1989; Meuser, 1994; Witt, 1997; Bergthaller et al., 1998) and the initial trials in this study, three sensitive steps were identified that affect the recoveries and purities of A-starch and gluten fractions in a HS-FWD process. They are the high-shear mixing of flour and water to give a flour– water dispersion, aging or maturation of the protein-rich phase ob-tained after centrifugal fractionation of a HS-FWD, and isolation and washing of the gluten fraction to remove contaminants. In this work, RSM regression equations generated from the experimental data had coefficients of determination (R2) of 0.95 for A-starch recovery and 0.65 for protein recovery without gluten aging (Table 1). The marginal goodness of fit of the regression equation for protein recovery is due mostly to the narrow range (66–72%) of protein recoveries. It was found that 14% of the RBS flour proteins were soluble in water at 35 °C. Therefore, the maximum protein recovery in the gluten fraction from the RBS flour would be approximately 86%.

Contour plots of the responses modeled in the high-shear mix-ing step of the HS-FWD test on the RBS flour are shown inFig. 5. The temperature of the highly sheared flour–water dispersion immediately after shearing varied from 26 to 43 °C (Fig. 5a). The temperature of the sheared dispersion increased with increasing water temperature and shear rate, but the water–flour ratio had little influence on the temperature of the sheared dispersion.

Recovery of the starch in the puriﬁed A-starch fraction depended mostly upon shear rate up to 6000 rpm, above which starch recovery reached a plateau (Fig. 5b). However, increasing the water temperature from 27 to 34 °C decreased the recovery of A-starch by about 6%, perhaps because of added swelling of starch granules leading to an increase in their buoyancy, or because of excessive strengthening of the gluten network at the elevated temperatures. The latter argument is supported by the increased recovery of pro-tein in the gluten fraction as the temperature of the water added to ﬂour was increased in the mixing step (Fig. 5c).

According toMeuser (1994), the gluten fibers that were created during high-shear mixing of flour and water aggregate into a por-ous network during subsequent centrifugation. The network is rel-atively buoyant and its elasticity during centrifugation is a crucial factor to the success of the centrifugal separation. During centrifu-gation individual starch granules are thought to stretch a micro-zone in the gluten network until it fails, which allows the granule to pass through and sediment. Upon the passage of the granule, the micro-zone of the network reforms connections between gluten threads thereby reestablishing the elasticity of the network. For an efficient separation of more than 50% of starch from the gluten, the elasticity of the network during centrifugation must be low en-ough to allow passage of starch granules in many micro-zones, yet large enough to resist collapse of the matrix on a macro-scale. Excessive aggregation of gluten proteins during the high-shear mixing step can increase protein recovery, yet at the same time de-crease starch recovery because the thicker gluten threads would interfere with passage of starch granules through the network.

Fig. 4. Scanning electron micrographs (SEMs) of (a) optimally mixed flour–water dough, (b) moderately sheared dough–water dispersion (MS-DWD), and (c) highly sheared flour–water dispersion (HS-FWD). The flour used in the experiments was the regional baking standard (RBS) flour, which is a composite of straight-grade flours milled from hard winter wheats.

(8)

The recovery of ﬂour protein in the gluten fraction was quite low (Fig. 5c), ranging from about 66–72%, as opposed to the theo-retical maximum of 86%. In the RSM experiment to model the high-shear mixing of ﬂour and water (Table 1), the protein-rich phase that was collected upon centrifugation of the HS-FWD was not aged or matured. Based on the assumption that aging of the glu-ten-rich phase might give more complete aggregation of proteins in the gluten fraction, another RSM optimization experiment was conducted as discussed below.

3.3. Modeling of the ‘‘gluten-aging’’ step in the HS-FWD wet-milling test; protein recovery in the gluten fraction

The recovery of starch in the A-starch fraction was over 80% when a HS-FWD was prepared from a highly sheared dispersion at a W/F ratio of 1.7, shear rate of 6000 rpm, and water tempera-ture of 25 °C (Fig. 5a). Under those conditions, 72% of ﬂour proteins were recovered in the gluten fraction by immediately processing the protein-rich phase in the Glutomatic system for 2.0 min (Fig. 5b). However, protein recoveries in the gluten fractions of a commercial high-shear process have been reported to range from 75% to 85%, often above 80% (Maijala, 1976; Dahlberg, 1978; Witt, 1997). In order to increase protein recovery in the HS-FWD

wet-milling test, the protein-rich phase was subjected to aging, a step that is reportedly used in the industrial high-shear process of Raisio (Kerkkonen et al., 1976; Maijala, 1976; Dahlberg, 1978).

Using the afore-mentioned conditions in the high-shear mixing step that gave high (>80%) recovery of starch in the A-starch fraction, an RSM model with an R2_{of 0.97 was developed based on the time}

and temperature conditions used to age the protein-rich phase ( Ta-ble 2andFig. 6). When the protein-rich phase was aged in added water at 30–45 °C for 10–30 min without agitation, 75% recovery of ﬂour proteins was achieved in the gluten fraction as opposed to 72% recovery without aging. At those aging conditions, the wet-glu-ten fraction was observed to be a cohesive with high elasticity, whereas it became too extensible at prolonged holding (>30 min) of the protein-rich phase in water at temperatures above 40 °C. Also, even a slight agitation of the protein-rich phase during aging moted disaggregation of the gluten particles and caused lower pro-tein recovery. Perhaps addition of extra water (100 mL) to the protein-rich phase (60 g, wb), which was done to control temper-ature, negatively impacted the gain in gluten recovery. The RSM modeling of the high-shear mixing step indicated that increasing the temperature of water added at the high-shear mixing step from 20 to 34 °C improved protein recovery (Fig. 5c) with a slight reduc-tion in starch recovery (Fig. 5b) in the A-starch fraction.

26 28 30 32 34 36 38 40 42 44 1,41,5 1,61,7 1,81,9 2,0 3000 4000 5000 6000 Te mpe ra ture of Highly She a re d Flour-Water Dispersion ( oC) W/F Ra tio Homoge nizer Speed (rpm) 34 oC 27 oC 20 oC 55 60 65 70 75 80 85 90 1,4 1,51,6 1,71,8 1,92,0 3000 4000 5000 6000 Starc h Re cov e ry in A -Sta rch Fraction (%) W/ F Rat io

Hom_{egenizer Speed} (rpm) 27 oC 34 oC 20 oC

(a)

(b)

65 66 67 68 69 70 71 72 73 1,4 1,51,6 1,71,8 1,92,0 3000 4000 5000 6000 Protein R e cov e ry in Gluten Fraction (% ) W/F Rati o Homoge nizer S_{peed (r} pm) 34 o_C 27 oC 20 oC

(c)

Fig. 5. Surface plots of (a) temperature of the highly sheared flour–water dispersion (HS-FWD) after shearing, (b) starch recovery in the A-starch fraction, and (c) protein recovery in the gluten fraction as determined through modeling of the high-shear mixing step of HS-FWD wet-milling test (RSM regression equations are given inTable 1). The flour used in the experiments was the regional baking standard (RBS) flour.

(9)

After setting the aging conditions of the protein-rich phase at 40 °C for 20 min, a series of experiments were conducted to deter-mine whether increasing water temperature from 25 to 40 °C in the high-shear mixing step would influence recovery of flour pro-tein in the gluten fraction (Fig. 7). During this series of experiments the W/F ratio was kept at 1.7, mixing speed at 6000 rpm, and mix-ing time at 2.0 min. As the water temperature was increased from 25 to 35 °C, the protein recovery increased from 75 to 79%. Fur-ther increasing water temperature to 40 °C did not significantly change the protein recovery (P > 0.05). The recovery of the A-starch fraction was slightly reduced (83% vs. 80%) as the temperature of the water added to the flour increased from 25 to 40 °C.

Witt (1997) showed in an industrial high-shear wet-milling process that aging (15 min) of a ﬂour–water dispersion prepared by a high-pressure homogenizer prior to fractionation in a

three-phase decanter centrifugation gave the best separation of the A-starch and the protein-rich phases. In this study, aging of the HS-FWD for 15 min as opposed to no-aging gave no statistically signiﬁcant differences (P > 0.05) in the recoveries of protein and starch in the gluten and starch fractions (data not shown).

Based on the above ﬁndings, the following process conditions for the bench-scale HS-FWD wet-milling test were chosen as opti-mum: W/F ratio of 1.7, water temperature of 35 °C, and homoge-nizer speed of 6000 rpm for 2.0 min in the high-shear mixing step; gluten-aging temperature and time of 40 °C and 20 min, respectively, in 100 mL of water; and a gluten-washing time of 2.0 min in the Glutomatic system ﬁtted with 88-

l

m opening sieves. Although variations in the W/F ratio had limited inﬂuence on the recoveries and purities of the fractions (Fig. 5), a W/F ratio of 1.7 was favored based on the ease of handling a dispersion dur-ing the homogenization and centrifugal fractionation steps.

3.4. Repeatability of the HS-FWD wet-milling test

The HS-FWD wet-milling test at the optimized conditions showed a high level of repeatability (Table 3). The coefficients of variation (CV) for the recoveries of the five wet-milling fractions and their purities ranged from 0.7 to 12.1%. The CV values for recov-ery of flour protein in the gluten fraction and recovrecov-ery of starch in the A-starch fraction, which are the two most important fractions in wet-milling of wheat flour, were 0.9 and 1.7%, respectively.

3.5. Comparison of three wet-milling tests

The wet-milling data from the HS-FWD test were compared to those from the MS-DWD and the Martin dough-washing tests ( Ta-bles 4 and 5), where all three tests were done on 65 g of the RBS flour. The distribution of flour solids in the wet-milling fractions of the three tests are given inTable 4. Total solids recovered in all wet-milling tests were above 98%. The Martin test gave the highest recoveries of solids in the gluten and A-starch fractions. The HS-FWD and MS-DWD tests gave similar recoveries of solids in their gluten fractions, but the recovery of solids in the A-starch fraction was lower in the MS-DWD. In the high- and medium-shear tests, the levels of flour solids in their water-solubles fractions were slightly elevated (Table 4).

The HS-FWD and MS-DWD tests gave similar recoveries (78– 79%) of protein in their gluten fractions with similar purities (83% protein content) (Table 5). However, the Martin dough-washing test, which involved minimal shear during processing, gave a higher recovery (86.7%) of protein in the gluten fraction con-taining 84.4% protein. The increased protein recovery in the gluten fraction of the Martin test likely occurred because much of the water-soluble proteins (14%) of the RBS flour became associated with the gluten fraction. On the other hand, the reduced recovery of protein in the HS-FWD and MS-DWD test may be explained by the high-shear forces employed in those processes during the preparation of the sheared dispersions prior to centrifugation, which caused the soluble-proteins to be dispersed and possibly some of the gluten particles to became so small in size that they were lost to the B-starch fraction. The insoluble polymeric protein contents of the HS-FWD and the RBS flour were comparable (40.9 vs. 42.2%), indicating that high-shear mixing of flour and water did not cause extensive solubilization of insoluble proteins.

A higher level of protein (4.6%) was recovered in the B-starch fraction of the HS-FWD test than in the MS-DWD (1.6%) and Martin dough-washing (2.0%) tests (Table 5). Also, approximately 30% more protein was recovered in the water-solubles fractions of the HS-FWD and MS-DWD tests than that in the Martin dough-washing method (14.8 vs. 9.5%). It was reported that almost all globulins and up to one-half of albumins might be present in the

B

C

A

Temperature of Water Added to RBS Flour in High-Shear Mixing Step of HS-FWD Test (°C)

25 30 35 40 Protein Recov ery in Gluten Fraction (% ) 70 72 74 76 78 80 82 B C C A

Fig. 7. Protein recovery in the gluten fraction as affected by the temperature of water added to flour in the high-shear mixing step of the HS-FWD wet-milling test (The high-shear mixing parameters were a W/F ratio of 1.7 and homogenizer speed of 6000 rpm for 2 min. The gluten-aging parameters were an aging temperature of 40 °C for 20 min. Recoveries with different letters indicate significant (P < 0.05) difference). The flour used in the experiment was the regional baking standard (RBS) flour. 72 73 74 75 76 10 20 30 40 25 30 35 40 Protein Recovery (%)

Aging Time (min) Aging Temperatu

re (oC)

Fig. 6. Surface plot of protein recovery in the gluten fraction upon aging of protein-rich phase as determined through modeling of the gluten-aging step of the HS-FWD wet-milling test (RSM regression equation is given inTable 2). The ﬂour used in the experiment was the regional baking standard (RBS) ﬂour.

(10)

gluten fraction isolated by the Martin test, because of their physi-cal entrapment and disulﬁde bonding (Pence et al., 1956). Fane and Fell, as cited byRausch (2002), reported that 16% of total ﬂour pro-tein was not recovered in the gluten fraction isolated by the Martin process, which translated into a protein recovery of 84% in the glu-ten fraction.

In terms of the recovery of starch in the A-starch fraction, the Martin test was the best (89.4%), followed by the HS-FWD (80.3%) and MS-DWD (71.8%) tests (Table 5). In the Martin test, the A-starch fraction was obtained by centrifugation of the starch milk that contained both the A-starch and B-starch fractions that were washed from a dough mass with copious amounts of water, and therefore some B-starch that would otherwise be recovered in the B-starch fractions of the HS-FWD and MS-DWD tests was recovered in the A-starch fraction of the Martin test. The MS-DWD test gave signiﬁcantly lower A-starch recovery (71.8%) than the HS-FWD test (80.3%) because the centrifugal separation of the MS-DWD was not as sharp as for the HS-FWD. Large aggregates of gluten with entrapped starch granules occurred in the MS-DWD as shown inFig. 4. The A-starch fractions obtained by the three wet-milling tests all contained similarly low levels of contaminat-ing proteins (0.20–0.24%), which was within the industrially

acceptable range. The recoveries of dry-solids and starch in the B-starch fractions were inversely associated with those in the A-starch fractions (Tables 4 and 5).

The properties of starches and vital glutens isolated by the three wet-milling tests, which were published elsewhere (Sayaslan et al., 2010), were quite comparable. Damaged starch levels (1.6–2.1%, db) and RVA pasting properties of the A-starch fractions isolated by all wet-milling tests were found similar, indicating that starch was not damaged during the shear treatments in the HS-FWD or MS-DWD tests. Similarly, the breadmaking qualities of the vital glutens isolated by the wet-milling methods were determined to be comparable, suggesting that gluten proteins were not damaged nor altered during the shear treatments.

3.6. Application of the HS-FWD wet-milling test to wheat ﬂours with varying gluten strength

3.6.1. Wet-milling of hard wheat ﬂours by the HS-FWD wet-milling test

The HS-FWD wet-milling test at the optimized conditions (Fig. 2) was used to evaluate the wet-milling properties of ﬁve hard wheats with different mixing properties, namely Karl 92, Jagger,

Table 4

Recovery of flour solids in the five wet-milling fractions from the RBS wheat flour using the highly sheared flour–water dispersion (HS-FWD), moderately sheared dough–water dispersion (MS-DWD) and dough-washing (Martin) wet-milling testsa

. Wet-milling test Fractions (% of ﬂour solids)

Gluten A-starch B-starch Fiber Water-solubles Total

HS-FWD 12.9b 63.1b 13.2b 0.9b 8.7a 98.8a

MS-DWD 13.1b 56.4c 20.6a 0.9b 7.2b 98.3a

Dough-washing (Martin) 14.3a 70.3a 6.4c 1.2a 6.2c 98.4a

a_{All percentages are on dry-solids basis. Different letters in the same column indicate signiﬁcant (P < 0.05) difference.} Table 3

Repeatability of the highly sheared ﬂour–water dispersion (HS-FWD) wet-milling test on the RBS wheat ﬂour at optimized conditions.a

Experiment day Gluten fraction A-starch fraction B-starch fraction

Protein recovery (% of flour protein) Protein content (%, Nx5.7) Starch recovery (% of flour starch) Protein content (%, Nx5.7) Starch recovery (% of flour starch) Protein content (%, Nx5.7) 1 78.1 83.0 79.2 0.25 12.6 4.2 2 79.5 82.4 81.7 0.26 11.0 4.4 3 78.9 83.8 81.2 0.24 10.6 4.9 4 79.4 82.4 78.5 0.23 12.1 3.9 5 78.1 82.7 80.9 0.23 11.4 5.4 Mean 78.8 82.9 80.3 0.24 11.5 4.6 Std. deviation 0.7 0.6 1.4 0.02 0.8 0.6 Coefficient of variation (%) 0.9 0.7 1.7 6.4 6.7 12.1 a

All percentages are on dry-solids basis. Optimized conditions were as follows: W/F ratio of 1.7, water temperature of 35 °C and homogenizer speed of 6000 rpm for 2.0 min in the high-shear mixing step; gluten-aging temperature and time of 40 °C and 20 min, respectively, in 100 mL of water; and gluten-washing time of 2.0 min in the Glutomatic in two washing chambers ﬁtted with 88-lm opening sieves.

Table 5

Comparison of fractions from the optimized highly sheared ﬂour–water dispersion (HS-FWD) wet-milling test with moderately sheared dough–water dispersion (MS-DWD) and dough-washing (Martin) wet-milling testa

. Wet-milling

test

Gluten fraction A-starch fraction B-starch fraction

Protein recovery (% of flour protein) Protein content (%, Nx5.7) Starch content (%) Starch recovery (% of flour starch) Starch content (%) Protein content (%, Nx5.7) Starch recovery (% of flour starch) Starch content (%) Protein content (%, Nx5.7)

HS-FWD 78.8b 82.9a 11.0b 80.3b 95.8a 0.24a 11.5b 87.6a 4.6a

MS-DWD 78.0b 83.5a 15.5a 71.8c 94.3a 0.20b 18.0a 87.3a 1.6c

Dough-washing (Martin)

86.7a 84.4a 7.1c 89.4a 94.2a 0.20b 5.2c 80.9b 2.0b

a

(11)

2137, TAM 110 and Chisholm (Fig. 1). As compared to the RBS, Karl 92 had a much stronger dough-mixing property while Jagger had a similar dough-mixing proﬁle. Flours of 2137 and TAM 110 showed relatively weak dough-mixing properties. On the other hand, Chis-holm ﬂour had an unusually long mixing time (14 min) due prob-ably to its low protein content.

Fig. 8shows photographs taken after the centrifugation of the HS-FWD of the RBS and five hard wheat flours, plus those of two additional flours, prepared by the optimized HS-FWD wet-milling test. The levels of A-starch sediments and the sharpness of their demarcations from the protein-rich phases (Fig. 8) correlated pos-itively with the dough-mixing properties of the hard wheat flours (Fig. 1). Karl 92 and Chisholm flours showed the highest proportion of sedimented A-starch phase while 2137 and TAM 110 showed the lowest. The HS-FWD of Jagger had intermediate behavior. It ap-pears that when the HS-FWD from a flour having strong

dough-mixing characteristics is subjected to centrifugal force, the fine-meshed gluten network that forms in the centrifugal field probably is strong and resists collapsing. The resistance to collapse allows more of the large starch granules to pass through the mesh resulting in a high level of sedimented starch. In the case of the flours with weak dough-mixing characteristics, however, the glu-ten network in their HS-FWD apparently suffers some collapse in the centrifugal field, which inhibits sedimentation of the large granules to the bottom phase. The collapse of a weak gluten net-work also may explain the less clearly defined demarcation plane between the starch and the protein-rich phases (Fig. 8).

Table 6 summarizes wet-processing data for the hard wheat flours obtained by the HS-FWD wet-milling test. Recoveries of flour protein (78.8–82.2%) in the gluten fractions isolated from all hard wheat flours were comparable to that of the RBS flour (78.8%), ex-cept for the reduced protein recovery (71.8%) from Chisholm flour.

Fig. 8. Centrifugal separation of the highly sheared ﬂour–water dispersions (HS-FWD) of various wheat ﬂours that have different dough-mixing properties.

Table 6

Wet-milling of various wheat ﬂours with different dough-mixing properties by the highly sheared ﬂour–water dispersion (HS-FWD) wet-milling testa

. Wheat

ﬂour

Fractions

Gluten A-starch B-starch Fiber Water-solubles Total (%

of flour) Protein recovery (% of flour protein) Protein content (%, Nx5.7) Starch recovery (% of flour starch) Protein content (%, Nx5.7) Fraction recovery (% of flour) Protein content (%, Nx5.7) Fraction recovery (% of flour) Dry-solids recovery (% of flour)

RBS (control) 78.8c 82.9b 80.3bc 0.24b 13.2c 4.6abc 0.9d 8.7cd 98.8a

Karl 92 78.8c 86.9ab 82.5b 0.22b 11.8c 4.3bcd 0.7e 9.7bc 100.3a

Jagger 82.2a 73.0c 76.8c 0.23b 13.9bc 4.7ab 0.7e 10.2bc 100.1a

2137 81.3ab 70.0c 64.2d 0.23b 23.4a 2.6d 0.9d 8.7cd 98.5a

TAM 110 79.2bc 60.3d 66.6d 0.23b 18.1b 2.9cd 1.1cd 10.5ab 99.1a

Chisholm 71.8d 89.8a 89.0a 0.24b 12.6c 6.3a 1.4c 7.3d 99.6a

Soft 79.9abc 65.3cd 76.9c 0.27ab 13.8bc 3.2cd 1.8b 5.9e 98.9a

Coarsely ground 61.0d 77.5bc 74.9c 0.32a 17.7b 7.5a 5.4a 12.0a 99.6a

a

(12)

However, protein concentrations in the gluten fractions isolated from the weak flours were generally lower than those of the flours with strong dough-mixing characteristics. All flours gave A-starch fractions with similar purities (0.22–0.24% protein), but the recov-eries of starch in the A-starch fractions of 2137 (64.2%) and TAM 110 (66.6%) were much lower than from the other flours (76.8– 89.0%). Recoveries of starch in the B-starch fractions followed an inverse trend to the recoveries in the A-starch fractions for all flours. Similar recoveries (>98.5%) of flour solids summed in the five wet-milling fractions were achieved from all flours (Table 6).

The results indicate that the aggregation characteristics of the gluten proteins in a HS-FWD influence the fractionation of flour components in their wet-processing by a high-shear wet-milling process. Flours with strong dough-mixing characteristics appear to give strong porous gluten networks upon high-shear mixing, and such a dispersion maintains a porous gluten macrostructure without collapse for an extended period of time in a centrifugal field. Consequently, separation of the large starch granules from the protein-rich phase occurs more efficiently. The more starch that is removed in the first centrifugation step, the easier it is to isolate and purify the gluten fraction. It is well-known that the large starch (A-type) granules in most wheat flours account for about 70% of total starch.

3.6.2. Wet-milling of a soft wheat ﬂour by the HS-FWD wet-milling test

A soft red winter wheat (AT 91W) flour with 12.2% protein (14% mb) was wet-milled by the optimized HS-FWD test. The mixogram (Fig. 1) of the soft wheat flour indicated a mellow gluten strength in dough-mixing, typical of soft wheat flours. When the HS-FWD of the soft wheat flour was centrifugally fractionated, separation of the phases was comparable to that of the RBS flour (Fig. 8). The soft wheat flour gave similar wet-milling results to the RBS flour (Table 6), except for the reduced purity of the gluten fraction. The gluten fraction was probably contaminated with starch and pentosans, as evidenced by the reduced recoveries of starch and solids in the A-starch and water-solubles fractions, respectively. The wet gluten of the soft wheat was also observed to stick to the Glutomatic sieve during washing, indicating weaker gluten characteristics typical of soft wheats.

3.6.3. Wet-milling of a coarsely ground wheat ﬂour by the HS-FWD wet-milling test

In the industrial wet-processing of wheat flour, a short-flow dry-milling process to produce coarsely ground flour is desirable to reduce mechanical damage to starch granules and to reduce the cost of dry-milling (Maningat and Bassi, 1999). A flour sample produced by a short-flow dry-milling process (78.0% flour yield from wheat) with coarse particles was wet-milled by the HS-FWD wet-milling test. The coarsely ground flour had a higher level of bran contamination than the straight-grade flours as indicated by its high ash content (0.77%, 14% mb).

The coarsely ground flour produced a mixogram comparable to that of the RBS flour (Fig. 1). When the highly sheared dispersion of the coarsely ground flour was centrifuged, a thick layer of fiber/ bran fraction was observed between the bottom starch phase and the slower sedimenting protein-rich phase (Fig. 8). A reduced vol-ume of starch was separated from the dispersion of the coarse flour as compared to that seen for the RBS flour (Fig. 8). Although total recoveries of flour solids in the five wet-milling fractions were sim-ilar, the coarsely ground flour gave significantly lower recoveries of protein in the gluten fraction and of starch in the A-starch fraction (Table 6). Also, the purities of those fractions were significantly re-duced. Not surprisingly the fiber fraction isolated from the coarsely ground flour was six times higher than that from the RBS flour due to the higher bran content of the coarsely ground flour. The

reduction in the recovery of protein from the coarsely ground flour might be caused at least in part by its relatively high content of nongluten proteins. It is well-known that as the flour yield in-creases in the dry-milling so does the level of nongluten proteins from the aleurone and bran layers. Those contaminating proteins lack the aggregation trait of endosperm gluten proteins. Most of those nongluten proteins are water-soluble, and as seen inTable 6, they became part of the water-solubles fraction rather than the gluten fraction during wet-processing. Another reason for the re-duced recovery of protein is that the high-shear mixing step was inadequate to completely disaggregate some of the large flour par-ticles into free starch granules and gluten threads. The reduction in the recovery of starch in the A-starch fraction from the coarse flour was attributed to its high level of bran fiber. Bran particles, ob-served to be swollen and bulky in the HS-FWD, likely interfered with the sedimentation of starch granules during the centrifugal fractionation step. In an industrial high-shear process that uses coarsely ground flour, the flour–water slurry should be sieved prior to high-shear mixing, which was practiced in the pilot-scale HD process ofMeuser (1994)starting with whole-wheat flour.

4. Conclusions

A bench-scale wet-milling test starting with a highly sheared flour–water dispersion (HS-FWD) was devised. The test was con-ducted on 65 g (db) of flour and yielded five wet-milling fractions: gluten, A-starch, B-starch, fiber, and water-solubles. The equip-ment required for the test includes a rotor–stator mixing homoge-nizer, a vibrating wet-sieve set, a floor-model centrifuge, and a gluten-washing instrument. The optimized HS-FWD wet-milling test differentiated the wet-milling qualities of several hard, soft, and coarsely ground wheat flours with varying dough-mixing properties. All flours gave quantitative recoveries of solids in the five fractions of the HS-FWD test. However, the A-starch fractions, all of which contained 0.2–0.3% protein, accounted for 64–89% recovery of flour starch, and the gluten fractions, containing 60– 90% protein, accounted for 61–81% recovery of flour protein.

References

AACC, 2000. Approved methods of the american association of cereal chemists (Approved Methods 08-01, 10-10B, 30-25, 38-10, 44-15A, 54-40A, 13, 76-31). American Association of Cereal Chemists, St. Paul, MN.

Anderson, R.A., 1967. Manufacture of wheat starch. In: Whistler, R.L., Paschall, E.F. (Eds.), Starch: Chemistry and Technology, vol. II. Academic Press, New York, pp. 53–63.

Anderson, R.A., 1974. Wet-processing of wheat ﬂour. In: Inglett, G.E. (Ed.), Wheat: Production and Utilization. AVI Publishing, Westport, CT, pp. 355–365. Bean, S.R., Lyne, R.K., Tilley, K.A., Chung, O.K., Lookhart, G.L., 1998. A rapid method

for quantitation of insoluble polymeric proteins in ﬂour. Cereal Chemistry 75, 374–379.

Bergthaller, W., Lindhauer, M.G., Zwingelberg, H., 1998. Variety testing of wheat for starch production using small scale process. In: Praznik, W., Huber, A. (Eds.), Contributions of the Fourth International Workshop on Carbohydrates as Organic Raw Materials. WUV-Universitätsverlag, Vienna, Austria, pp. 163–176. Bergthaller, W.J., 1997. New uses of wheat gluten and non-starch wheat components. In: Steele, J.L., Chung, O.K. (Eds.), Proceedings of International Wheat Quality Conference. Grain Industry Alliance, Manhattan, KS, pp. 285– 301.

Czuchajowska, Z., Pomeranz, Y., 1993. Protein concentrate and prime starch from wheat ﬂours. Cereal Chemistry 70, 701–706.

Czuchajowska, Z. & Pomeranz, Y. (1995). Process for fractionating wheat ﬂours to obtain protein concentrates and prime starch. U.S. Patent, 5439,526. Dahlberg, B.I., 1978. A new process for the industrial production of wheat starch and

wheat gluten. Starch/Stärke 30, 8–12.

Fellers, D.A., 1973. Fractionation of wheat into major components. In: Pomeranz, Y. (Ed.), Industrial Uses of Cereals. American Association of Cereal Chemists, St. Paul, MN, pp. 207–228.

Finney, K.F., Shogren, M.D., 1972. A ten-gram mixograph for determining and predicting functional properties of wheat ﬂours. Baker’s Digest 46(2), 32–35, 38–42, 77.

Godon, B., Leblanc, M.-P., Popineau, Y., 1983. A small scale device for wheat gluten separation. Qualitas Plantarum/Plant Foods for Human Nutrition 33, 161–168.

(13)

Hamer, R.J., Weegels, P.L., Marseille, J.P., Kelfkens, M., 1989. A study of the factors affecting the separation of wheat ﬂour into starch and gluten. In: Pomeranz, Y. (Ed.), Wheat Is Unique. American Association of Cereal Chemists, St. Paul, MN, pp. 467–477.

Kerkkonen, H.K., Laine, K.M.J., Alanen, M.A., Renner, H.V., 1976. Method of Separating Gluten From Wheat Flour. US Patent, 3951,938.

Knight, J.W., Olson, R.M., 1984. Wheat starch: production, modiﬁcation, and uses. In: Whistler, R.L., BeMiller, J.N., Paschall, E.F. (Eds.), Starch Chemistry and Technology, second ed. Academic Press, New York, pp. 491–506.

Lindhauer, M.G., 1997. Selection of wheats and their dry-milling prior to wet-milling. In: Steele, J.L., Chung, O.K. (Eds.), Proceedings of International Wheat Quality Conference. Grain Industry Alliance, Manhattan, KS, pp. 217–230. Lindhauer, M.G., Bergthaller, W.P., 2002. Characteristics for the assessment of starch

extractability from wheat. In: Yuryev, V.P., Cesaro, A., Bergthaller, W.J. (Eds.), Starch and Starch Containing Origins: Structures, Properties and New Technologies. Novo Science Publishers, New York, pp. 375–382.

Maijala, M., 1976. Environment and product gain in new wet wheat process. Food Engineering 48, 73–75.

Maningat, C.C., Bassi, S.D., 1999. Wheat starch production. In: Tumbleson, M., Yang, P., Eckhoff, S. (Eds.), Proceedings of International Starch Technology Conference. University of Illinois, Urbana, IL, pp. 26–40.

Maningat, C.C., Seib, P.A., 1997. Update on wheat starch and its use. In: Steele, J.L., Chung, O.K. (Eds.), Proceedings of International Wheat Quality Conference. Grain Industry Alliance, Manhattan, KS, pp. 261–284.

Maningat, C.C., Bassi, S., Hesser, J.M., 1994. Wheat gluten in food and non-food systems. Technical Bulletin, XVI, 6. American Institute of Baking, Manhattan, KS. Meuser, F., 1994. Wheat utilization for the production of starch, gluten and extruded products. In: Bushuk, W., Rasper, V.F. (Eds.), Wheat: Production, Properties and Quality. Chapman & Hall, New York, pp. 179–204.

Meuser, F., Althoff, F., Huster, H., 1989. Developments in the extraction of starch and gluten from wheat ﬂour and wheat kernels. In: Pomeranz, Y. (Ed.), Wheat Is Unique. American Association of Cereal Chemists, St. Paul, MN, pp. 479–499. Pence, J.W., Mecham, D.K., Olcott, H.S., 1956. Review of proteins in wheat ﬂour.

Journal of Agricultural and Food Chemistry 4, 712–715.

Rausch, K.D., 2002. Front end to backpipe: membrane technology in the starch processing industry. Starch/Stärke 54, 273–284.

Robertson, G.H., Cao, T.K., 1998a. Substitution of concentrated ethanol for water in the laboratory washing fraction of protein and starch from hydrated wheat ﬂour. Cereal Chemistry 75, 508–513.

Robertson, G.H., Cao, T.K., 1998b. Methods for Separation of Wheat into Protein and Starch Fractions. US Patent, 5851,301.

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, 1334–1343.

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, 1344–1349.

Sayaslan, A., 2002. Bench-Scale Wet-Milling of Wheat Flour: Development of a Test to Fractionate a Highly Sheared Flour-Water Dispersion and Its Comparison with Fractionation by the Dough-Washing Test. Ph.D. Dissertation. Kansas State University, Manhattan, KS.

Sayaslan, A., 2004. Wet-milling of wheat ﬂour: Industrial processes and small-scale test methods. LWT – Food Science and Technology 37, 499–515.

Sayaslan, A., Seib, P.A., Chung, O.K., 2006. Wet-milling properties of waxy wheat ﬂours by two laboratory methods. Journal of Food Engineering 72, 167–178. Sayaslan, A., Seib, P.A., Chung, O.K., 2010. Properties of starch and vital gluten

isolated from wheat ﬂour by three different wet-milling methods. Journal of Applied Biological Sciences 4 (2), 51–56.

Schafer, R., 1975. Yield determination in laboratory. A method for quantitative ﬂour evaluation in wheat starch industries. Starch/Stärke 27, 257–262.

Seib, P.A., 1994. Wheat starch: isolation, structure and properties. Oyo Toshitsu Kagaki 41, 49–69.

Van Der Borght, A., Goesaert, H., Veraverbeke, W.S., Delcour, J.A., 2005. Fractionation of wheat and wheat ﬂour into starch and gluten: overview of the main processes and the factors involved. Journal of Cereal Science 41, 221–237. Verberne, P., Zwitserloot, W., 1978. A new hydrocyclone process for the separation

of starch and gluten from wheat ﬂour. Starch/Stärke 30, 337–338.

Walker, C.E., Parkhurst, A.M., 1984. Response surface analysis of bake-lab data with a personal computer. Cereal Foods World 29, 662–666.

Wang, M., 2003. Effect of Pentosans on Gluten Formation and Properties. Ph.D. Dissertation. Wageningen University, Wageningen.

Wang, M., Hamer, R.J., Vliet, T.V., Oudgenoeg, G., 2002. Interaction of water extractable pentosans with gluten protein: effect on dough properties and gluten quality. Journal of Cereal Science 36, 25–37.

Wang, M., Hamer, R.J., Vliet, T.V., Gruppen, H., Marseille, H., Weegels, P.L., 2003a. Effect of water unextractable solids on gluten formation and properties: mechanistic considerations. Journal of Cereal Science 37, 55–64.

Wang, M., Oudgenoeg, G., Vliet, T.V., Hamer, R.J., 2003b. Interaction of water unextractable solids with gluten protein: Effect on dough properties and gluten quality. Journal of Cereal Science 38, 95–104.

Weegels, P.L., Marseille, J.P., Hamer, R.J., 1988. Small scale separation of wheat ﬂour in starch and gluten. Starch/Stärke 40, 342–346.

Witt, W., 1997. Modern method of separating the components of wheat. In: Steele, J.L., Chung, O.K. (Eds.), Proceedings of International Wheat Quality Conference. Grain Industry Alliance, Manhattan, KS, pp. 213–248.

Zwitserloot, W.R.M., 1989. New integrated process for starch, gluten production. World Grain, June, 12–15.