ARTICLE IN PRESS

0733-5210/$ - se

doi:10.1016/j.jc

Abbreviations

thermal analysi

Gordon–Taylor

TVP, texturised�Correspond

YO91 1XY, UK

E-mail addr

(A. Arrachid).

Journal of Cereal Science 45 (2007) 275–284

www.elsevier.com/locate/yjcrs

Relationship between the glass transition temperature and themelt flow behavior for gluten, casein and soya

Carlos Bengoecheaa, Abdessamad Arrachidb,�, Antonio Guerreroa,Sandra E. Hillb, John R. Mitchellb

aDepartamento de Ingenierı́a Quı́mica, Universidad de Sevilla, c/ P. Garcia Gonzalez 1, 41012 Sevilla, SpainbDivision of Food Sciences, School of Biosciences, University of Nottingham, Sutton Bonington Campus, Loughborough LE12 5RD, UK

Received 29 January 2006; received in revised form 11 June 2006; accepted 25 August 2006

Abstract

The effects of moisture content (25–45% wwb) and temperature (75–120 1C) on the viscosity of gluten, soya and rennet casein systems

was studied using a capillary rheometer. An attempt was made to relate the viscosities to the glass transition temperature measured by

differential scanning calorimetry, dynamic mechanical thermal analysis and the phase transition analyzer. The temperature where the

material flowed was also determined by the latter technique. All three-protein systems showed shear and extension thinning. Over the

shear rate range investigated (�1–103 s�1), gluten had a substantially lower viscosity than the other two proteins, although the difference

was less pronounced at the highest temperature studied. This low viscosity is reflected by lower values of the glass transition temperature,

the melt flow temperature and the dynamic moduli E0 and E00 in the rubbery state. The results are discussed in terms of the structure and

heat induced changes for the three proteins and their relevance to food processing considered.

r 2006 Elsevier Ltd. All rights reserved.

Keywords: Glass transition; Viscosity; Rheology; Extrusion; Differential scanning calorimetry; Dynamic mechanical thermal analysis; Phase transition

analyzer; Plasticization; Protein

1. Introduction

The rheological behavior of protein ‘‘melts’’ at relativelylow water contents (�15–50% wwb) is important forextrusion processing to produce food products such astexturised vegetable proteins (TVP). In addition, rheologi-cal behavior of gluten at low water contents is one factorinfluencing baking performance. There have been extensivestudies on the viscosity of biopolymer melts using bothpressure capillary rheometry and on-line extrusion rhe-ometers (Bhattacharya, 1993; Breuillet et al., 2002; Fujio

e front matter r 2006 Elsevier Ltd. All rights reserved.

s.2006.08.011

: C–K, Couchman–Karasz; DMTA, dynamic mechanical

s; DSC, differential scanning calorimetry; G–T,

; PTA, phase transition analyzer; SPI, soya protein isolate;

vegetable protein

ing author. PTC NESTLE YORK, Haxby Road, York

. Tel.: +44 1904 60 31 16; fax: +44 1904 60 48 87.

ess: Abdessamad.Arrachid@rdyo.nestle.com

et al., 1991; Singh and Smith, 1999; Zhang et al., 1998). Thevariables generally studied are temperature, shear rate andwater content (Colonna et al., 1989; Harper, 1981)although time dependent changes such as protein associa-tions have sometimes been taken into account (Morganet al., 1989; Remsen and Clark, 1978).More recently there has been an increasing interest in the

glassy state in foods (Blanshard and Lillford, 1993). Theglass transition temperature (Tg) has been measuredextensively for both carbohydrate and protein dominatedsystems, generally using differential scanning calorimetry(DSC) or dynamic rheological methods. Of particularinterest for food applications has been the plasticizing roleof water, sugars, polyols and other low molecular weightadditives (Kalichevsky et al., 1992a, b, 1993; Morales andKokini, 1997; Pommet et al., 2003; Roos, 1995; Zhanget al., 2005).A technique that has been recently developed to provide

information both about the glass transition temperature

ARTICLE IN PRESSC. Bengoechea et al. / Journal of Cereal Science 45 (2007) 275–284276

and the melt rheology of biopolymer systems, particularlywithin the context of extrusion processing, is the phasetransition analyser (PTA) (Plattner et al., 2001). In thePTA a piston applies pressure to material initially in aclosed chamber and the temperature is increased. The glasstransition temperature is associated with the point wherethe material softens, giving rise to increased movement ofthe piston. A flow temperature, Tf, can also be determinedby replacing the closed die at the base of the chamber byone containing a small hole. Tf is taken as the temperaturewhere the material flows as evidenced by a continuousmovement of the piston at a constant temperature. Formaize these two temperatures have been shown to differ forcrops grown in different locations and have been related toextrusion behavior (Plattner et al., 2001).

In this study, three protein systems were examined:gluten, soya and casein. All three are used in extrusionprocesses, and gluten and soya are also important in bakedproducts. Large amounts of information can be found inthe literature about their structure and their properties inextrusion (Bhattacharya and Hanna, 1986; Chen et al.,1978; Damodaran and Paraf, 1997; Jao et al., 1978;Kitabatake and Doi, 1992). Gluten is an amorphousmixture of polypeptides that can be divided according totheir functionality into monomeric and polymeric poly-peptides. Monomeric polypeptides consist mainly ofstorage proteins called gliadins. They can aggregate bynon-covalent interactions and contribute to the viscosityand extensibility of gluten. The vast majority of polymericpolypeptides are found in the glutenin fraction. Theiramino acid compositions are similar to the gliadins. Theyare additionally stabilized by disulfide bonds, and tend tocontribute to the elasticity and strain hardening behaviorof gluten (MacRitchie and Lafiandra, 1997; Weegels et al.,1996). Soya consists of mainly 7S and 11S globulins. The7S globulin is a trimer with a molecular weight around150–200 kD (Fukushima, 1991). The 11S globulin iscomposed by six subunits and has a reported molecularweight of 300–400 kD (Fukushima, 1991; Pearson, 1982).Soya is used to produce TVP (Zhang et al., 2001) and isfunctionally involved in gelation and emulsification (Utsu-mi et al., 1997). Casein is the principal protein in bovinemilk and consists of four components, as1- (38%), as2-(10%), b- (36%) and k-casein (13%), which have molecularweights in the range of 19–26 kD and vary in hydro-phobicity (Dalgleish, 1997). In milk, they are arranged in amicelle. In rennet-coagulated casein, micellar aggregationby hydrophobic bonding is thought to precede coagulation.Rennet casein performs well in dry spinning processeswhere it is extruded at temperatures of 70 1C and above(Visser, 1988).

One objective of the work described in this paper is todetermine if more conventional methods for determining aglass transition temperature, DSC, dynamic mechanicalthermal analysis (DMTA) and melt rheology (pressurecapillary rheometry) are consistent with the informationobtained from the PTA. A second objective was to obtain

further information about the relationship between proteintype, the glass transition temperature and the meltrheology.

2. Materials and methods

2.1. Materials

Rennet casein was obtained from Kerry Foods Ltd(High Protein Milk Extract, UK), gluten from RIBA S.A.(Glutenflor Supervital, Barcelona, Spain), and soya proteinisolate (SPI) from Protein Technologies International(SUPRO 500E, Leper, Belgium).

2.2. Methods

2.2.1. Sample preparation

2.2.1.1. Capillary rheometry. Samples were prepared byadding the required amount of water to the powder andmixing with a Kenwood mixer (KenWood Mixer KMC500, KenWood) to ensure homogeneity. Gluten forms aviscoelastic dough when mixed with water; so to prepare ahomogenous powder it was necessary to freeze thehydrated material in liquid nitrogen, prior to milling it toa powder using a Knifeter 1095 Sample Mill (Foss Tecator,Hoganas, Sweden). For all three proteins water contentwas measured after each experiment performed with thecapillary rheometer and was found to match the predictedwater content.

2.2.1.2. DSC and PTA. Samples were equilibrated overP2O5 or saturated salt solutions of MgCl2, Mg(NO3)2, KI,NaCl or KNO3 or distilled water, which producedequilibrium relative humidities (RH) of 0%, 32.8%,52.8%, 68.9%, 75.3%, 93.7% and 100%, respectively.Samples were equilibrated at room temperature for at leastseven days.

2.2.1.3. DMTA. Samples were prepared by hydrating to�17% water at 100% RH overnight and then pressing to athickness of 0.5–1mm in a mold under pressure�3.1� 103 kPa at a temperature between 70 and 90 1C.Samples were then cut into 20� 8� 1mm strips and storedover salt solutions of various relative humidities (as forDSC and PTA: Section 2.2.1.2) to obtain a variety of watercontents. Samples were stored for at least a week beforemeasurements were made. Following equilibration, thewater content was checked for agreement with predictedvalues and the measured values were used in theinterpretation of the results. Before the measurement, thesample was coated with silicone oil (Dow Corning, USA)to avoid water loss.

2.3. Water content

Water contents were obtained by drying to constantweight at 105 1C.

ARTICLE IN PRESSC. Bengoechea et al. / Journal of Cereal Science 45 (2007) 275–284 277

2.4. Capillary rheometry

A Rosand RH7 twin bore capillary rheometer (RosandFlowmaster RH7, Bohlin Instruments, UK) was used. Thesample was driven simultaneously through a capillary die(l ¼ 32mm, + ¼ 0:5mm, y ¼ 901) as well as an orifice die(+ ¼ 0:5mm, y ¼ 901). The samples (25%, 30%, 35%,40% and 45% (wwb) of water content) were loaded in thebarrels, allowed to equilibrate for 10min and extruded atdifferent temperatures (65, 75, 85, 95, 120 1C). Theextrusion was carried out at ram speeds of 2, 5.63, 15.9,44.7, 126, 355 and 1000mm/s. The pressure was recordedas a function of the piston speed.

The data were fitted to the power law equation as used inprevious studies (Colonna et al., 1989; Harper, 1981; Singhand Smith, 1999):

s ¼ m_gn, (1)

where s is the shear stress, _g is the shear rate, n is the flowbehavior index and m is the consistency coefficient(Brydson, 1981).

The instrument software obtains the shear viscosity ðZ ¼s=_gcÞ from the Rabinowitsch corrected wall shear rate:

_gc ¼3nþ 1

4n

� �_g, (2)

where _g the wall shear rate for a Newtonian fluid is givenby

_g ¼4Q

pr3

� �, (3)

Q is the volumetric flow rate and r the capillary radius.The wall shear stress is given by

s ¼ðPL � P0Þr

2LL

, (4)

where PL is the pressure drop across the long die, lengthLL, and P0 the pressure drop across a zero length dieobtained from the pressure drop across the orifice die.

The Cogswell method (Cogswell, 1972) is used to obtainan extensional viscosity ðl ¼ se=�Þ, where the extensionalstress is given by

se ¼3ðnþ 1Þ

8P0 (5)

and the extensional strain rate by

� ¼4

3

Z_g2

ðnþ 1ÞP0. (6)

2.5. DSC

Calorimetric measurements were performed using aPerkin Elmer DSC-7 (Perkin Elmer, UK). An emptystainless steel pan was used in the reference holder. Thesample was heated first time at 10 1C/min from �60 to180 1C. It was then cooled to �60 1C at 50 1C/min and

heated a second time at 10 1C/min to 180 1C. Finally thesample was cooled to 20 1C at 50 1C/min. The glasstransition temperature, Tg, was determined as the tem-perature midpoint of the heat capacity change observedduring the second run.

2.6. DSC data fitting

To predict the Tg of mixtures from the Tg’s of thecomponents the empirical Gordon–Taylor (Gordon andTaylor, 1952) (G–T) equation was used:

Tg ¼W 1Tg1 þ KW 2Tg2

W 1 þ KW 2, (7)

where W is the mass fraction of each component, Tg is theglass transition temperature in Kelvin, and K is a systemconstant inversely proportional to the plasticizing effect ofthe diluent (1) on the protein (2). This model is suitable forbicomponent systems (food component and water or otherplasticizer) and has proven to be useful to fit DSC data forstarch, maltose, maltodextrins (Roos and Karel, 1991) andfor cereal proteins (de Graaf et al., 1993; Kokini et al.,1995; Madeka and Kokini, 1996). For multicomponentsystems, the Couchman–Karasz (Couchman and Karasz,1978) (C–K) equation has been used:

Tg ¼W 1DCp1Tg1 þW 2DCp2Tg2

W 1DCp1 þW 2DCp2. (8)

This equation requires the value of DCp of water, whichis subject to considerable debate (Kalichevsky et al.,1992b). G–T equation is equivalent to the C–K equationif K ¼ DCp2=DCp1 (4), and DCp is the change in heatcapacity observed at Tg. The values used for Tg1 and DCp1

of water are 134K (�139 1C) and 1.94 J g�1K�1, respec-tively (Kalichevsky et al., 1993; Sugisaki et al., 1968).Values of DCp2 for the dry protein (casein, soya, gluten)were calculated from Eq. (3), assuming a value of1.94 J g�1K�1 for DCp1 for water.

2.7. PTA

The PTA is a closed-chamber capillary ‘‘rheometer’’.One of the main differences between PTA and a capillaryrheometer is that the latter extrudes at constant pistonspeed while the PTA performs at constant pressure.It consists of two sealed chambers, top and bottom,

separated by an interchangeable capillary die. The twochambers house electric heaters and contain a hollowcavity that allows a cooling fluid to be used. The pistons,mounted together as sidebars, are held in a fixed positionduring testing. Air cylinders, mounted to the bottom of thePTA, maintain constant pressure on the sample. A linear-displacement transducer measures the samples’ deforma-tion, compaction, and flow relative to initial sample height(Plattner et al., 2001). In this study the measurements of Tg

and Tf were carried out at a heating rate of 5 1C/min and anapplied pressure of 150 bar.

ARTICLE IN PRESS

0

1

2

3

4

5

6

7

8

9

10

0 20 40 60 80 100 120 140 160 180 200

Temperature (C)

Dis

pla

cem

en

t (m

m)

Casein

Soya

Gluten

Fig. 2. Displacement temperature plots from the phase transition analyser

indicating position of glass transition temperature. Moisture contents

were: soya (13.7%, wwb), casein (12.9%, wwb) and gluten (11.7%, wwb).

Applied pressure: 150 bar. Heating rate: 5 1C/min.

C. Bengoechea et al. / Journal of Cereal Science 45 (2007) 275–284278

2.8. Dynamic mechanical thermal analyzer

The dynamic mechanical thermal analyzer (RheometricScientific, Piscataway, USA) was used with the samplepresented in the single cantilever-bending mode at afrequency of 1Hz and a strain of 0.03% which was in thelinear region. The heating rate used was 3 1C/min. Theresults from two runs were averaged at each water content.

3. Results

3.1. Glass transition temperature

Fig. 1 displays the DSC response for the second DSCscan for the three proteins at a single, comparable moisturecontent. All three proteins showed an endothermic peakimmediately above the glass transition temperature in theDSC first scans. This was particularly pronounced forgluten (first run scan shown in Fig. 1). This endothermicpeak has been reported previously for a number of proteinsincluding gluten subunits (Castelli et al., 2000) and isprobably due to physical aging (enthalpy relaxation) aneffect well known for synthetic polymers (Drozdov, 2001;Hay, 1993; Toufeili et al., 2002). The midpoint of the heatcapacity change from the reheat was taken as the glasstransition temperature. This change was less clear for soyathan for casein and gluten. Morales and Kokini (1997)have reported that at, for example, a water content of 15%the glass transition temperature of the 11S globulin isabout 40 1C higher than the 7S globulin. Since the soyaisolate used in this work is a mixture of these two proteins,this may explain the breadth of the transition.

Fig. 2 shows the piston displacement as a function oftemperature from the PTA. As the material begins tosoften, the piston movement increases more rapidly withtemperature. The glass transition temperature was taken asthe temperature where the derivative of the displacementtemperature plot was a maximum. An interesting feature ofthe displacement temperature response is the higher initial

0 20 40 60 80 100 120 140 160 180

Temperature (C)

Heat

flo

w -

En

do

therm

ic Gluten

Casein

Soya

Gluten 1st run

Fig. 1. DSC response for the three proteins during the reheat. The first

run for gluten only is represented. Moisture contents were: soya (13.7%,

wwb), casein (12.9%, wwb) and gluten (11.7%, wwb). Heating rate: 10 1C/

min.

bulk density and the lower degree of compression as thetemperature increases for gluten compared with the othertwo proteins. This may suggest that even in the glassy state,gluten is more deformable than the other proteins but suchan interpretation needs to be treated with caution as thegluten samples were comminuted in a different way(freezing followed by subsequent milling).From DMTA data, as a function of temperature the

glass transition temperature can be taken as the tempera-ture where a decrease occurs in E0 or the maximum in tan d( ¼ E00/E0) or E00. In this work, the E00 maximum was takenas the glass transition temperature since this falls inbetween the other two temperatures. Fig. 3 shows thetan d, E0 and E00 responses for all three proteins atcomparable water contents. A peak appears in the DMTAscans around �60 1C independent of water content (Fig. 3).Kalichevsky et al. (1992b) also reported this low tempera-ture transition. They stated that this may be due to theonset of short range motions, whereas the glass transition isthe onset of main chain motion. Di Gioia et al. (1999)indicated two weak relaxations at �65 and �25 1C in theDMTA scans attributed to secondary relaxation ofproteins. There is considerable evidence from a range oftechniques for a glassy transition in all polypeptide watersystems at a temperature around 200K (Ringe and Petsko,2003) and it is possible that this mechanical relaxation isanother manifestation of this. Whereas this mobilitytransition may be relevant to the low temperature storageof biological materials, it does not influence the mainrheological responses of the material. Gluten shows ahigher decrease in the storage modulus �102 Pa and also aconsiderably higher tan d peak than found for the othertwo proteins in the glass transition region. Our results areconsistent with data of Kalichevsky et al. (1993) whocompared the tan d and E0 responses for gluten, casein,caseinate, ovalbumin and gelatin around the glass transi-tion. They also found a fall in E0 of approximately 102 Paand a tan d maximum close to 0.75 at the glass transition ofgluten at a water content of 15%. These data also

ARTICLE IN PRESS

1.E+06

1.E+07

1.E+08

1.E+09

-100 -50 0 50 100 150 200

Temperature (C)

E" (

Pa)

Soya

Casein

Gluten

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

-100 -50 0 50 100 150 200

Temperature (C)

Tan

delt

a

Socv

Casein

Gluten

1.E+06

1.E+07

1.E+08

1.E+09

1.E+10

-100 -50 0 50 100 150 200

Temperature (C)

E' (P

a)

Soya

Casein

Gluten

Fig. 3. E0 and E00 and tan d responses from DMTA for the three proteins.

Moisture contents were soya (13.7%, wwb), casein (12.9%, wwb) and

gluten (11.7%, wwb). Frequency: 1Hz. Heating rate: 3 1C/min.

0

20

40

60

80

100

120

140

160

180

200

0 5 10 15 20 25 30

Moisture content (%, wwb)

Tem

pera

ture

(C

)

DMTA

DSC

PTA

CASEIN

0

20

40

60

80

100

120

140

160

180

200

0 5 10 15 20 25 30

Moisture content (%, wwb)

Tem

pera

ture

(C

)

DMTA

DSC

PTA

SOYA

0

20

40

60

80

100

120

140

160

180

200

0 5 10 15 20 25 30

Moisture content (%, wwb)

Tem

pera

ture

(C

)

DMTA

DSC

PTA

GLUTEN

a

b

c

Fig. 4. Comparison of the effect of the water content on Tg values

measured using DSC, DMTA and PTA for casein, soya and gluten.

C. Bengoechea et al. / Journal of Cereal Science 45 (2007) 275–284 279

suggested a much lower glass transition temperature forgluten than obtained for casein and caseinate, as also foundin our work. Of the proteins in this investigation, onlygelatin gave a comparable E0 decrease to gluten andshowed a large tan d peak (�0.9). The mechanicallymeasured glass transition temperature for gelatin as judgedfrom the temperature of the tan d peak was higher thangluten but lower than the casein systems. This relativelylarge fall in modulus of 102 Pa has also been reported fordry gluten by Pouplin et al. (1999), who found a value ofthe tan d maximum of 0.75 compared with 0.8 shown in

Fig. 3. It is appropriate to add, however, that Pouplin et al.(1999) reported that the tan d maximum varied with watercontent and also reported an absolute value for E0 whichwas 102 Pa lower than shown in Fig. 3 for the samplecontaining 11.7% water.Fig. 4 compares the Tg values for the three proteins from

the different techniques. DMTA gave the highest Tg andDSC the lowest. Kalichevsky et al. (1992a) also indicated ahigher value for the glass transition temperature obtainedfrom DMTA at 1Hz than the glass transition temperaturefrom DSC, though the differences were smaller than in ourstudy. Of particular interest are the values given by the

ARTICLE IN PRESS

0

50

100

150

200

250

0 5 10 15 20 25 30

Moisture content (%, wwb)

Tem

pera

ture

(C

)

Tg

Tf

CASEIN

0

50

100

150

200

250

0 5 10 15 20 25 30

Moisture content (%, wwb)

Tem

pera

ture

(C

)

Tg

TfSOYA

150

200

250

ratu

re (

C)

Tg

TfGLUTEN

C. Bengoechea et al. / Journal of Cereal Science 45 (2007) 275–284280

PTA. This measures a temperature where the modulus ofthe powder decreases and might be expected to relate mostclosely to the temperature where the storage modulus E0

falls. The lower temperature for the PTA determined Tg

compared with the DMTA value taken from the E00

maximum is consistent with this. It should be appreciatedthat the glass transition is a kinetic phenomenon which willdepend not only on the criteria used to but also on the timescale of measurement.

All three techniques show that gluten has the lowest glasstransition temperature at all moisture contents, and caseinthe highest. Table 1 shows the results from the fitting of theG–T to the DSC values for Tg. The glass transitionpredicted for dry gluten using this approach is in goodagreement, within experimental error, with the valuesreported in the literature, (433K (Hoseney et al., 1986);435K (Kalichevsky et al., 1992b); 412–448K (for differentsubunits of gluten) (Noel et al., 1995; Sartor and Johari,1996); 423K (Cherian and Chinachoti, 1996); 460K(Pouplin et al., 1999); 448K (Micard et al., 2001); 401K(Toufeili et al., 2002). Similar agreement was found withthe Tg values for dry casein by Kalichevsky et al. (1992b)(Tg ¼ 417K) and Mizuno et al. (1999) (Tg ¼ 408K)) andfor soya (387 (7S)–433 (11S) (Morales and Kokini, 1997);423–473K (by DMA) (Ogale et al., 2000)).

The very high Tg in the absence of plasticizers comparedwith most synthetic polymers (Aklonis and MacKnight,1983; Ferry, 1970) could be explained by the presence of asignificant incidence of polymer-polymer interactions, by ahigh density of hydrogen bonding, or by ionic orhydrophobic interactions (Pouplin et al., 1999).

0

50

100

0 5 10 15 20 25 30

Moisture content (%, wwb)

Tem

pe

Fig. 5. Effect of the moisture content on Tg and the melt flow temperature

Tf for casein, soya and gluten. Tg and Tf were obtained at a pressure of

150 bar and a heating rate of 5 1C/min.

3.2. Flow temperature (Tf)

Fig. 5 compares Tf and Tg both measured at a pressureof 150 bar for the three proteins at different moisturecontents. It can be seen that gluten which has the lowest Tg,also starts to flow at much lower temperatures than theother two proteins. For example, at a water content of 15%Tf for gluten is �75 1C compared with �140 1C for soyaand �120 1C for casein. Another important feature is theconstant difference in temperature between Tg and Tf forthe three proteins irrespective of the water content. Thisdifference is �37 1C for gluten, �39 1C for casein and�75 1C for soya. This result suggests a strong link betweenthe glass to rubber transition and the ability of the proteinsto flow. The higher difference between Tg and Tf found for

Table 1

Gordon-Taylor and Couchman-Karasz fitting parameters obtained for

gluten, soya and casein

Protein Tg (K) K DCp [k ¼ DCp2=DCp1] (J g�1K�1)

Gluten 434.9 4.9370.61 0.394

Soya 445.0 4.4270.79 0.439

Casein 479.4 4.9770.73 0.391

soya can be related to the longer transition observed in theDSC plot as well as in the DMTA (E0).

3.3. Capillary viscosity

Fig. 6 shows the results for the shear and extensional ratedependence of the viscosity from the capillary rheometer atthe highest and the lowest water contents (45% (wwb) and25% (wwb)) and at temperatures of 120 and 75 1C. All theproteins show shear and extensional thinning behaviors forall the conditions studied. The extensional viscosities aremuch higher than the shear viscosities with a Trouton ratioof 10–70 at a common deformation rate of vicinity of1.101 s�1. In all cases, the shear viscosity at the lower shearrates is much lower for gluten than it is for the other two

ARTICLE IN PRESS

1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

1.E+06

1.E+07

1.E+08

1.E+09

1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04

Extensional or shear rate (1/s)

Ex

ten

sio

na

l a

nd

Sh

ea

r v

isc

os

ity

(Pa

.s)

Casein

Soya

Gluten1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

1.E+06

1.E+07

1.E+08

1.E+09

1.E+10

1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04

Extensional or shear rate (1/s)

Ex

ten

sio

na

l o

r s

he

ar

vis

co

sit

y

(Pa

.s)

Casein

Soya

Gluten

1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

1.E+06

1.E+07

1.E+08

1.E+09

1.E+10

1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04

Extensional or shear rate (1/s)

Ex

ten

sio

na

l o

r s

he

ar

vis

co

sit

y

(Pa

.s)

Casein

Soya

Gluten1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

1.E+06

1.E+07

1.E+08

1.E+09

1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04

Extensional or shear rate (1/s)

Ex

ten

sio

na

l o

r s

he

ar

vis

co

sit

y

(Pa

.s)

Casein

Soya

Gluten

Fig. 6. Extensional and shear viscosity of the three proteins at different temperatures and moisture contents. Extensional viscosities are represented by the

highest values. Conditions were: (a) 75 1C–25% (wwb); (b) 75 1C–45% (wwb); (c) 120 1C–25% (wwb); (d) 120 1C–45% (wwb).

Table 2

Model fitting parameters for the capillary shear data for gluten, soya and

casein

Protein K0 n DE (kJ/mol) a R2

Gluten 2340 0.45 5.70 4.96 0.78

Soya 7.86 0.34 30.3 5.83 0.91

Casein 0.399 0.27 29.8 8.04 0.95

C. Bengoechea et al. / Journal of Cereal Science 45 (2007) 275–284 281

proteins. The difference is greatest for the conditionsclosest to the glass transition temperature (75 1C and 25%moisture content) with gluten having a shear viscositybetween one and two orders of magnitude lower than soyaand casein. This is consistent with the Tf values measuredwith the PTA which at 20% water is about 50 1C lowerthan the other two proteins. The glass transition tempera-ture is also lower for gluten than the other two proteins aspreviously discussed. Soya always gives the highest valuesof extensional viscosities compared with the two otherproteins. The difference is highest for the low temperature.

Some general appreciation of the effect of deformationrate, moisture content and temperature on viscosity can beobtained by fitting an equation of the form (Colonna et al.,1989):

Z ¼ k0 expDE=RT exp�aMC _gn�1, (9)

where Z is the viscosity, k0, a constant, DE, the activationenergy, MC, the moisture content of the sample, R, theuniversal gas constant, T, the absolute temperature, _g inthis case is the shear or extensional rate, n, the flowbehavior index, and a, a constant, to the data obtained atall five temperatures and moisture contents. Since dena-turation and time dependent crosslinking reactions willoccur at the higher temperatures it should be appreciated

that the constants obtained will be dependent on thetime–temperature history and thus should be regarded onlyas an empirical representation of the data set. Asmentioned in the introduction, models have been proposedto take crosslinking into account (Morgan et al., 1989;Remsen and Clark, 1978), but to include this requires amuch more complex and extensive experimental procedure.Tables 2 and 3 show the fitted values for the constants inEq. (9) for the shear and extensional viscosities, respec-tively. The values of n show that the materials appear toshow more extension thinning than shear thinning as it canbe clearly seen from Fig. 6. The values also show that, ingeneral, the gluten shear viscosity is less dependent ontemperature and moisture content than the other proteins,whereas the temperature dependence of the extensionalviscosity is much higher for soya.

ARTICLE IN PRESS

Table 3

Model fitting parameters for the capillary extensional data for casein,

gluten and soya

Protein K0 n DE (kJ/mol) a R2

Gluten 22000 0.17 3.02 6.78 0.95

Soya 0.050 0.16 42.5 7.38 0.94

Casein 3790 0.21 14.8 8.69 0.88

C. Bengoechea et al. / Journal of Cereal Science 45 (2007) 275–284282

4. Discussion

Results showed the difficulty to relate viscosity (shearand extensional) behavior to glass transition temperature.The extensional viscosity is related to the pressure drop atthe entrance to the capillary of the capillary rheometer andwill be influenced by features such as the resistance tomolecular elongation and the extent of molecular entangle-ment between these molecules. Extensional viscosity doesnot appear to relate to the glass transition temperature inthe way that shear viscosity does. It might be suggestedthat a higher then expected Trouton ratio for gluten couldbe related to the strain hardening properties of gluten inextension (Breuillet et al., 2002), whereas in comparisonwith casein, soya possess more ordered globular structuresresisting orientation.

Of particular interest is the very low shear viscosityfound for the gluten melt compared with the other twoproteins. The zero shear viscosity is related to the stressrelaxation modulus G(t) by the expression:

Z ¼Z 10

tGðtÞd ln t, (10)

where the frequency dependence of G0 mirrors the timedependence of the stress relaxation modulus G(t) (Ferry,1970) (high frequencies will be equivalent to short timesand the extensional modulus EðtÞ ¼ 3GðtÞ). If it is assumedthat time temperature superposition holds, then a lowviscosity relates both to a low Tg, a low value of G* or G0 inthe rubbery state, a narrow transition and a narrowrubbery region in terms of time or temperature. For acompletely amorphous high molecular weight polymerwith increasing temperature, there will be a glass to rubbertransition which is evidenced by a fall in the elasticmodulus of about 103 Pa to a value governed by the densityof non-specific and ill defined chain entanglements(Aklonis and MacKnight, 1983; Ferry, 1970). In additionto the fall in modulus, the glass transition is evidenced by apeak in tan d. The height of the tan d peak relatesquantitatively to the volume fraction of the relaxing phaseand for a completely amorphous polymer the height of thepeak is �1.0, and the breadth of this transition is typically10–20 1C (Wetton et al., 1986). The extensive studies ofwater plasticised biopolymers have invariably shownsmaller and broader transitions than would be expectedfor an ideal high molecular weight synthetic polymer. Forexample, the fall in modulus is often only an order ofmagnitude and the height of the tan d peak is less than 0.5

(Kalichevsky et al., 1993). The relationship betweenpolysaccharide structure and a calorimetrically determinedTg has been extensively discussed by Bizot et al. (1997) whoconsider the increase in Tg with various types of chainimmobilization. For a 1-4 linked polysaccharide chains, itis not possible to observe a glass transition calorimetricallypossibly because of specific inter residue hydrogen bonds(Gidley et al., 1993). Other interactions which contribute toimmobilize the polymer e.g. crystallization, the presence ofordered secondary structure (e.g., b-sheets or a-helices inproteins), covalent crosslinks e.g., disulfide bonds, andother interchain non-covalent interactions sometimescalled hyperentanglements (Morris et al., 1981) would beexpected to increase Tg, and if they are preserved attemperatures above Tg, they would reduce the size andincrease the width of the glass transition. All these effectswill increase the viscosity.The low shear viscosity of gluten relative to the two

other proteins studied along with the higher value of tan dand low moduli in the rubbery region, suggests that theimmobilizing interactions are lower in this protein systemthan in most other biopolymers i.e., it is closer to an idealsynthetic polymer. The reason for this is not clear. Thegreater hydrophobicity as shown by the lack of watersolubility may be a factor. Another possibility could bedifferences in the content of proline+hydroxyprolineresidues in the proteins. These residues are not compatiblewith most ordered secondary structures (a-helix, b-sheet)and, therefore, will be another factor contributing to anexpanded disordered structure rather than one that iscompact and ordered (Mohammed et al., 2000). Glutenand gelatin, which also show a large fall (�102 Pa) in thestorage modulus at the glass transition (Kalichevsky et al.,1993), both have a higher praline/hydroxyproline contentthan soya and casein (Mohammed et al., 2000).It should be appreciated that the range of factors

contributing to the immobilization of the polypeptidechain makes a detailed interpretation difficult. Forexample, Mizuno et al. (2000) demonstrated an increasein Tg on crosslinking casein with microbial transglutami-nase as would be expected, but a counter-intuitive decreasefor soya. The latter was attributed to a reduction in b sheetsecondary structure associated with the crosslinking reac-tion. An important complicating factor is that heating willresult in additional crosslinking. This is indicated byexotherms observed in DSC studies of gluten at low watercontents and discussed in detail by Sartor and Johari(1996). This behavior of gluten protein could be explainedby disulfide/thiol exchange and thiol oxidation that takeplace upon heating, leading to an increase of the covalentcoupling by disulfide bonds of gluten proteins (Morel et al.,2000; Schofield et al., 1983; Weegels and Hamer, 1998).Crosslinks involving tyrosine have also been suggested(Tilley et al., 2001).Protein crosslinking, which in general would be expected

to increase with temperatue and water content, willincrease the viscosity of the protein melt.

ARTICLE IN PRESSC. Bengoechea et al. / Journal of Cereal Science 45 (2007) 275–284 283

The relative contributions of the glass transition andcrosslinking reactions can be understood qualitatively fromEq. (10). If the time dependence of the relaxation modulusis considered, crosslinking will increase both the width andheight of the rubbery plateau and an increase in the glasstransition temperature will increase the width of the glassyzone. Both these effects will increase viscosity, which is anintegral of the stress relaxation modulus over time. Theobservation that under all conditions gluten has a lowershear viscosity than the other two proteins, particularly atlow shear rates where Eq. (10) is likely to be most relevant,suggests that what dominates its low viscosity is the low Tg.The low values for DE and a for gluten compared with theother two proteins is probably due to crosslinking reactionsbecoming more dominant at higher temperatures andwater contents resulting in a lower decrease in viscositywith increasing temperature and water content. Cross-linking could explain the increase in the modulus (E0) seenfor casein and gluten at high temperatures (Fig. 3) thoughwater loss might also be a factor.

5. Conclusions

The PTA gives values for the glass transition tempera-ture and also the temperature of flow which are consistentwith the information obtained from DSC, DMTA andcapillary rheometry. Gluten gives a much lower shearviscosity than soya and casein which is partly related to alower Tg and the involvement of more of the protein in themobilization occurring above Tg. This is shown not only bytransition techniques (DSC, DMTA, capillary rheometry)but also by the measurements of Tg and Tf using the PTA.This may be expected to have implications not only forextrusion processing but for the performance of gluten inbaked products.

References

Aklonis, J.J., MacKnight, W.J., 1983. Introduction to polymer viscoelas-

ticity. Wiley, New-York.

Bhattacharya, M., 1993. Slit rheometer studies of wheat-flour dough.

Journal of Texture Studies 24, 391–409.

Bhattacharya, M., Hanna, M.A., 1986. Viscosity modelling of dough in

extrusion. Journal of Food Technology 21, 167–174.

Bizot, H., Le Bail, P., Leroux, B., Davy, J., Roger, P., Buleon, A., 1997.

Calorimetric evaluation of the glass transition in hydrated linear and

branched polyanhydroglucose compounds. Carbohydrate Polymers

32, 33–50.

Blanshard, J.M.V., Lillford, P.J., 1993. The Glassy State in Foods.

Nottingham University Press, Nottingham.

Breuillet, C., Yildiz, E., Cuq, B., Kokini, J.L., 2002. Study of the

anomalous capillary Bagley factor behavior of three types of wheat

flour doughs at two moisture contents. Journal of Texture Studies 33,

315–340.

Brydson, J.A., 1981. Flow Properties of Polymer Melts. George Godwin,

London.

Castelli, F., Gilbert, S.M., Caruso, S., Maccarone, D.F., Fisichella, S.,

2000. Thermoanalytical characterisation of high molecular weight

glutenin subunits. Water effect on their glass transition. Thermo-

chimica Acta 346, 153–160.

Chen, A.H., Jao, Y.C., Larkin, J.W., Goldstein, W.E., 1978. Rheological

model of soy dough in extrusion. Journal of Food Process Engineering

2, 337–342.

Cherian, G., Chinachoti, P., 1996. H and O Nuclear magnetic resonance

study of water in gluten in the glassy and rubbery state. Cereal

Chemistry 73, 618–624.

Cogswell, F.N., 1972. Measuring extensional rheology of polymer melts.

Transactions of the Society of Rheology 16, 383–403.

Colonna, P., Tayeb, J., Mercier, C., 1989. Extrusion cooking of starch and

starchy products. In: Mercier, C., Linko, P., Harper, J.M. (Eds.),

Extrusion Cooking. American Association of Cereal Chemists, St.

Paul, MN, p. 247.

Couchman, P.R., Karasz, F.E., 1978. A classical thermodynamic

discussion of the effect of composition on glass transition tempera-

tures. Macromolecules 11, 117–119.

Dalgleish, D.G., 1997. Structure-function relationship of caseins. In:

Damodaran, S., Paraf, A. (Eds.), Food Proteins and their Applica-

tions. Marcel Dekker Inc., New York, Basel, Hong Kong,

pp. 199–223.

Damodaran, S., Paraf, A., 1997. Food Proteins and their Applications.

Marcel Dekker Inc., New York, Basel, Hong Kong, p. 681.

de Graaf, E., Madeka, H., Cocero, A.M., Kokini, J.L., 1993. Determina-

tion of the effect of moisture on gliadin glass transition using

mechanical spectrometry and differential scanning calorimetry. Bio-

technology Progress 9, 210–213.

Di Gioia, L., Cuq, B., Guilbert, S., 1999. Thermal properties of corn

gluten meal and its proteic components. International Journal of

Biological Macromolecules 24, 341–350.

Drozdov, A.D., 2001. Physical aging in amorphous polymers: comparison

of observations in calorimetric and mechanical tests. European

Polymer Journal 37, 1379–1389.

Ferry, J.D., 1970. Viscoelastic Properties of Polymers. Wiley, New York,

London, Sydney, Toronto.

Fujio, Y., Hayashi, N., Hayakawa, I., 1991. Effect of moisture content on

flow behaviour of molten soy-protein isolate under an elevated

temperature. International Journal of Food Science and Technology

26, 45–51.

Fukushima, D., 1991. Structures of plant storage proteins and their

functions. Food Reviews International 7, 353–381.

Gidley, M.J., Cooke, D., Ward-Smith, S., 1993. Low moisture poly-

saccharide systems: thermal and spectroscopic aspects. In: Blanshard,

J.M.V., Lillford, P.J. (Eds.), The Glassy State in Foods. Nottingham

University Press, Nottingham, pp. 303–315.

Gordon, M., Taylor, J.S., 1952. Ideal copolymers and the 2nd-order

transitions of synthetic rubbers.1. Non-crystalline copolymers. Journal

of Applied Chemistry 2, 493–500.

Harper, J.M., 1981. Extrusion in foods. CRC, Boca Raton, USA.

Hay, J.N., 1993. Physical ageing in polymer blends. In: Blanshard, J.M.V.,

Lillford, P.J. (Eds.), The Glassy State in Foods. Nottingham

University Press, Nottingham, pp. 269–279.

Hoseney, R.C., Zeleznak, K.J., Lai, L.S., 1986. Wheat gluten: a glassy

polymer. Cereal Chemistry 63, 285–286.

Jao, Y.C., Chen, A.H., Lewandowski, D., Irwin, W.E., 1978. Engineering

analysis of soy dough rheology in extrusion. Journal of Food Process

Engineering 2, 97–112.

Kalichevsky, M.T., Jaroszkiewicz, E.M., Ablett, S., Blanshard, J.M.V.,

Lillford, P.J., 1992a. The glass transition of amylopectin measured by

DSC, DMTA and NMR. Carbohydrate Polymers 18, 77–88.

Kalichevsky, M.T., Jaroszkiewicz, E.M., Blanshard, J.M.V., 1992b. Glass

transition of gluten. 1. Gluten and gluten sugar mixtures. International

Journal of Biological Macromolecules 14, 257–266.

Kalichevsky, M.T., Blanshard, J.M.V., Marsh, R.D.L., 1993. Application

of mechanical spectroscopy to the study of glassy biopolymers and

related systems. In: Blanshard, J.M.V., Lillford, P.J. (Eds.), The

Glassy State in Foods. Nottingham University Press, Nottingham,

pp. 133–156.

Kitabatake, N., Doi, E., 1992. Denaturation and texturization of food

protein by extrusion cooking. In: Kokini, J.L., Ho, C.-T., Karwe,

ARTICLE IN PRESSC. Bengoechea et al. / Journal of Cereal Science 45 (2007) 275–284284

M.V. (Eds.), Food Extrusion Science and Technology. Marcel Dekker

Inc., New York, Basel, Hong Kong, pp. 361–372.

Kokini, J.L., Cocero, A.M., Madeka, H., 1995. State diagrams help

predict rheology of cereal proteins. Food Technology 49 (74), 76–82.

MacRitchie, F., Lafiandra, D., 1997. Structure-function relationships of

wheat proteins. In: Damodaran, S., Paraf, A. (Eds.), Food Protein and

their Applications. Marcel Dekker Inc., New York, Basel, Hong

Kong, pp. 293–324.

Madeka, H., Kokini, J.L., 1996. Effect of glass transition and crosslinking

on rheological properties of zein: development of a preliminary state

diagram. Cereal Chemistry 73, 433–438.

Micard, V., Morel, M.-H., Bonicel, J., Guilbert, S., 2001. Thermal

properties of raw and processed wheat gluten in relation with protein

aggregation. Polymer 42, 477–485.

Mizuno, A., Mitsuiki, M., Motoki, M., 1999. Glass transition temperature

of casein as affected by transglutaminase. Journal of Food Science 64,

769–799.

Mizuno, A., Mitsuiki, M., Motoki, M., Ebisawa, K., Suzuki, E., 2000.

Relationship between the glass transition of soy protein and molecular

structure. Journal of Agricultural and Food Chemistry 48, 3292–3297.

Mohammed, Z.H., Hill, S.E., Mitchell, J.R., 2000. Covalent crosslinking

in heated protein systems. Journal of Food Science 65, 221–226.

Morales, A., Kokini, J.L., 1997. Glass transition of soy globulins using

differential scanning calorimetry and mechanical spectrometry. Bio-

technology Progress 13, 624–629.

Morel, M.H., Bonicel, J., Micard, V., Guilbert, S., 2000. Protein

insolubilization and thiol oxidation in sulfite-treated wheat gluten

films during aging at various temperatures and relative humidities.

Journal of Agricultural and Food Chemistry 48, 186–192.

Morgan, R.G., Steffe, J.F., Ofoli, Y., 1989. A generalized viscosity model

for extrusion of protein doughs. Journal of Food Process Engineering

11, 55–78.

Morris, E.R., Cutler, A.N., Ross-Murphy, S.B., Rees, D.A., Price, J.,

1981. Concentration and shear rate dependence of viscosity in random

coil polysaccharide solutions. Carbohydrate Polymers 1, 5–21.

Noel, T.R., Parker, R., Ring, S.G., Tatham, A.S., 1995. The glass-

transition behavior of wheat gluten proteins. International Journal of

Biological Macromolecules 17, 81–85.

Ogale, A.A., Cunningham, P., Dawson, P.L., Acton, J.C., 2000.

Viscoelastic, thermal and microstructural characterisation of soy

protein isolate films. Journal of Food Science 65, 672–679.

Pearson, A.M., 1982. Soy proteins. In: Hudson, B.J.W. (Ed.), Develop-

ments in Food Proteins. Elsevier Applied Science, London, pp. 67–108.

Plattner, B., Strahm, B., Rausch, K., 2001. ASAE Annual International

Meeting, Sacramento, California, USA; ASAE, St. Joseph, MI; Vol.

ASAE Meeting Paper Number 01-6067.

Pommet, M., Redl, A., Morel, M.H., Guilbert, S., 2003. Study of wheat

gluten plasticization with fatty acids. Polymer 44, 115–122.

Pouplin, M., Redl, A., Gontard, N., 1999. Glass transition of wheat gluten

pasticized with water, glycerol or sorbitol. Journal of Agricultural and

Food Chemistry 47, 538–543.

Remsen, C.H., Clark, J.P., 1978. A viscosity model for a cooking dough.

Journal of Food Process Engineering 2, 39–64.

Ringe, D., Petsko, G.A., 2003. The ‘‘glass transition’’ in protein dynamics:

what it is, why it occurs, and how to exploit it. Biophysical Chemistry

105, 667–680.

Roos, Y.H., 1995. Characterization of food polymers using state

diagrams. Journal of Food Engineering 24, 339–360.

Roos, Y.H., Karel, M., 1991. Water and molecular weight effects on the

glass transition in amorphous carbohydrates and carbohydrate

solutions. Journal of Food Science 56, 1676–1681.

Sartor, G., Johari, G.P., 1996. Polymerization of a vegetable protein,

wheat gluten, and the glass-softening transition of its dry and reacted

state. Journal of Physical Chemistry 100, 19692–19701.

Schofield, J.D., Bottomley, R.C., Timms, M.F., Booth, M.R., 1983. The

effect of heat on wheat gluten and the involvement of sulfhydryl-

disulphide interchange reactions. Journal of Cereal Science 1, 241–253.

Singh, N., Smith, A.C., 1999. Rheological behaviour of different cereals

using capillary rheometry. Journal of Food Engineering 39, 203–209.

Sugisaki, M., Suga, H., Seki, S., 1968. Calorimetric study of glassy state. 4.

Heat capacities of glassy water and cubic ice. Bulletin of the Chemical

Society of Japan 41, 2591.

Tilley, K.A., Benjamin, R.E., Bargorogoza, R.E., Okot-Kotber, B.M.,

Prakash, O., Kwen, H., 2001. Tyrosine cross-links molecular basis of

gluten structure and function. Journal of Agricultural and Food

Chemistry 49, 2627–2632.

Toufeili, I., Lambert, I.A., Kokini, J.L., 2002. Effect of glass transition

and cross-linking on rheological properties of gluten: development of a

preliminary state diagram. Cereal Chemistry 79, 138–142.

Utsumi, S., Matsumura, Y., Mori, T., 1997. Structure-function relation-

ships of soy proteins. In: Damodaran, S., Paraf, A. (Eds.), Food

Proteins and their Applications. Marcel Dekker Inc., New York,

Basel, Hong Kong, pp. 257–292.

Visser, J., 1988. Dry spinning of mil proteins. In: Blanshard, J.M.V.,

Mitchell, J.R. (Eds.), Food Structure—its Creation and Evaluation.

Butterworths, London, pp. 197–218.

Weegels, P.L., Hamer, R.J., 1998. Temperature-induced changes of wheat

products. In: Hamer, R.J., Hoseney, R.C. (Eds.), Interactions: the

Keys to Cereal Quality. American Association of Cereal Chemist Inc.,

St Paul, pp. 95–130.

Weegels, P.L., Hamer, R.J., Schofield, J.D., 1996. Functional properties of

wheat glutenin. Journal of Cereal Science 23, 1–17.

Wetton, R.E., Morton, M.R., Rowe, A.M., 1986. Comparison of DMTA

and DETA. American Laboratory 18, 96.

Zhang, W., Hoseney, R.C., Faubion, J.M., 1998. Capillary rheometer of

corn endosperm: glass transition, flow properties, and melting of

starch. Cereal Chemistry 75, 863–867.

Zhang, J., Mungara, P., Jane, J., 2001. Mechanical and thermal properties

of extruded soy protein sheets. Polymer 42, 2569–2578.

Zhang, X.Q., Burgar, I., Do, M.D., Lourbakos, E., 2005. Intermolecular

interactions and phase structures of plasticized wheat proteins

materials. Biomacromolecules 6, 1661–1671.