at SciVerse ScienceDirect

Food Hydrocolloids 33 (2013) 215e224

Contents lists available

Food Hydrocolloids

journal homepage: www.elsevier .com/locate/ foodhyd

The breakdown properties of heat-set whey protein emulsion gels inthe human mouth

Qing Guo a, Aiqian Ye a,*, Mita Lad a, Douglas Dalgleish b, Harjinder Singh a

aRiddet Institute, Massey University, Private Bag 11 222, Palmerston North 4442, New ZealandbDepartment of Food Science, University of Guelph, Canada

a r t i c l e i n f o

Article history:Received 29 November 2012Accepted 17 March 2013

Keywords:Emulsion gelMasticationMechanical propertiesBreakdown propertiesCorrelation

* Corresponding author. Tel.: þ64 6 350 5072; fax:E-mail addresses: A.M.Ye@massey.ac.nz, a.m.ye@m

0268-005X/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.foodhyd.2013.03.008

a b s t r a c t

Differently structured whey protein emulsion gels were formed by heating at different concentrations ofNaCl. The formation of gels was monitored by oscillatory rheometry. The large deformation propertiesrelevant to breakdown properties in the human mouth were measured by a uniaxial compression testand fracture wedge set test using a texture analyzer. A panel of 8 subjects was used to examine the in-mouth behaviours of gels including mastication parameters, degree of fragmentation and oil dropletrelease. The results showed that in general the gel hardness increased with increasing NaCl concen-tration. The gels containing 10/25 and 100/200 mM NaCl were characterized as being soft and hard,respectively. These soft and hard gels had different breakdown patterns in the mouth. On the other hand,sensory experiments showed the gel with 10 mM NaCl needed a significantly lower number of chewingcycles (19.4 � 2.1) compared with gels with higher NaCl. The values of median size of particles inmasticated gels containing 10, 25, 100 and 200 mM NaCl were about 4.00, 2.85, 1.05 and 0.95 mm,respectively, which suggested that higher hardness led to greater fragmentation in the human mouth.The fragmentation of the gel was highly correlated with functions of the mechanical properties. Therewas no obvious coalescence of the oil droplets during oral processing and only very few oil droplets werereleased from protein matrix during mastication.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Whey proteins are extensively used in the food industry as in-gredients as they have good functionality and high nutritionalvalue. The aggregation and cross-linking of whey proteinmoleculesinto three-dimensional solid-like networks (‘gels’) is one of themost important functional properties for developing microstruc-ture with desirable textural attributes (Dickinson & Hong, 1997).The type of heat-set whey protein gel is mainly determined by thestrength of the electrostatic interactions (Mehalebi, Nicolai, &Durand, 2008). At neutral pH and low ionic strength, a ‘fine-stranded’ gel is formed by the headetail association of short strands(primarily aggregates of whey protein molecules 50 and 10 nm inlength and diameter, respectively). Increasing ionic strength leadsto decreasing intermolecular repulsion, and coarser, particulategels are formed from the coagulation of the of small, dense ag-gregates (formed by random association of short strands) of whichthe size could be on a micrometre scale (Ikeda & Morris, 2002;

þ64 6 350 5655.assey.ac.nz (A. Ye).

All rights reserved.

Mehalebi et al., 2008; Pouzot, Nicolai, Visschers, & Weijers, 2005).These structural differences are reflected in the macroscopicproperties of the gel. Fine-stranded gels are rubbery/soft with asmall quantity of liquid release under compression, whereas par-ticulate gels are hard/brittle with large amounts of liquid release(Chantrapornchai & McClements, 2002; Ikeda, Foegeding, &Hagiwara, 1999).

Soft solid systems containing dispersed particles (e.g. emulsion-filled gels) have received recent interest due to their potential ap-plications in a range of industries from pharmaceutical andcosmetic to food formulations (Chen & Dickinson, 1999). Differenttypes of foods can be categorized as soft solid systems, including setyogurt, fresh cheese, puddings, dairy desserts and sausages, whichall have a structure referred to as emulsion gel (Rosa, Sala, Van Vliet,& Van De Velde, 2006). Recently, there have been a number ofstudies conducted to understand the structural formation andphysical properties of gelled whey protein emulsions (Boutin,Giroux, Paquin, & Britten, 2007; Chen & Dickinson, 1998, 1999;Dickinson & Chen, 1999; Dickinson & Yamamoto, 1996;McClements, Monahan, & Kinsella, 1993; Sok Line, Remondetto, &Subirade, 2005; Ye & Taylor, 2009). These studies suggest thatemulsified fat (active filler, inactive filler, varying size, different oil

Q. Guo et al. / Food Hydrocolloids 33 (2013) 215e224216

content) and the process of gel formation (for example, heating,acidification and enzyme action) strongly influence the structureand the rheological properties of the emulsion gel. The finalstrength of gels is determined by the magnitude of interactions(hydrogen bond, hydrophobic interaction, electrostatic interaction,covalent bond, etc.) between their different structural elements(protein matrix e protein matrix, protein matrix e oil droplet andoil droplet e oil droplet). Recently, fracture properties of wheyprotein emulsion gels have drawn increasing interest (Abhyankar,Mulvihill, & Auty, 2011; Rosa et al., 2006; Sala, de Wijk, van deVelde, & van Aken, 2008; Sala, van Vliet, Cohen Stuart, Aken, &van de Velde, 2009; Sala, van Vliet, Cohen Stuart, van de Velde, &van Aken, 2009). Gwartney, Larick, and Foegeding (2004) studiedthe sensory texture characteristics and large deformation proper-ties of whey protein emulsion gels. They found that stranded gelswith smooth and slippery surfaces broke down into large and non-homogeneous particles, whereas particulate gels broke down intosmall and homogenous particles. However, to better understandthe relationship between the structures and breakdown propertiesin the human mouth of whey protein emulsion gels, further workregarding to the large deformation/fracture properties is necessary.

When a foodproduct is consumed, it is exposed to awide range ofphysical (e.g. mechanical breakdown and temperature) andbiochemical (e.g. dilution effect, pH, enzymes, salts, mucin) pro-cesses in the human mouth, especially for semi-solid or solid foods(Foegeding et al., 2011; Sarkar, Goh, & Singh, 2009). Within the hu-man mouth, a bolus is formed by the mechanical action of chewingand biochemical processing by enzymes and proteins in the saliva,enabling safe swallowing of the food. Thus, oral processing leads toparticle size reduction, and contributes to the taste and texture offoods (Foegeding et al., 2011). And mechanical breakdown (frag-mentation) is a core part of oral processing (Chen, 2009).

The degree of fragmentation of a food product is criticallydependent on the structural properties of the food consumed(Agrawal, Lucas, Prinz, & Bruce, 1997; Lucas, Prinz, Agrawal, &Bruce, 2002). In general, harder foods require more chewing cy-cles and masticatory force, and probably lead to a higher degree offragmentation during mastication (Chen, 2009). However, foodswith the same hardness may have totally different degrees offragmentation, demonstrating the importance of the originalstructures of the food on the fragmentation process (Jalabert-Malbos, Mishellany-Dutour, Woda, & Peyron, 2007; Peyron,Mishellany, & Woda, 2004). Agrawal et al. (1997) and Lucas et al.(2002) found the cracking of food in the mouth is highly corre-lated with mechanical property index: toughness (R) and Young’smodulus (E). The toughness is defined as the energy consumed ingrowing a crack of a given area. Young’s modulus represents therigidity of the food material. These authors first developed the foodproperty indices including (R/E)0.5, (RE)0.5 and R to reflect the de-gree of fragmentation for different foods. However, there areobvious limitations in their studies, since they did not consider theeffect of saliva and temperature on mastication and the degree ofbreakdown of food was obtained only after one chew. Therefore,further study is needed on the relationship between the true de-gree of breakdown of food in the humanmouth and themechanicalproperties of the food.

Whey protein emulsion gels have a simple composition and canmimic real food which has an emulsion structure. By varying theconditions under which they are made, they provide a systematicway to precisely manipulate the structures and offer a good modelfor investigating the structural factors affecting the breakdown offoodmaterials during oral processing. The aims of this workwere toexplore the effect of different gel structures induced by NaCl on thephysical/mechanical properties of heat-set whey protein emulsiongels in large deformation and their breakdown properties in the

human mouth. The relationship between physical/mechanicalproperties and breakdown properties was also explored.

2. Materials and methods

2.1. Materials

Whey protein isolate with 90% protein content (WPI 895) waspurchased from Fonterra Co-operative Group Limited. Soybean oilwas purchased from the local market. Milli-Q grade water (Milli-pore Corp., Bedford, MA, USA) was used for all experiments, exceptwhenmaking samples for themastication studies, when food gradede-ionized water was used. Salt in the form of sodium chloride formastication experiments was food grade (purity > 99%). All thechemical reagents used in this study were of analytical grade andused without further modification unless otherwise stated.

2.2. Preparation of emulsions

Whey protein isolate (WPI) solutions were prepared by addingthe powder to Milli-Q water and stirring for 6 h at roomtemperature.

Pre-emulsions containing 10 wt% WPI and 20 wt% soybean oilwere prepared using a high-speed mechanical mixer (L5M, Silver-son, Massachusetts, USA) at 9000 rpm for 3 min. These pre-emulsions were then homogenized using 4 passes through a two-stage valve homogenizer (APV 2000; Albertslund, Denmark)operated at pressure of 300 bars for the first-stage and 25 bars forthe second-stage. The volume weighted mean diameter (D4,3) ofstock emulsions was around 0.48 mm. The stock emulsions, apartfrom the food grade emulsions, contained 0.02% sodium azide andwere stored at 4 �C until further use.

2.3. Rheological measurements

Dynamic oscillatory rheological measurements were carried outusing a controlled stress rheometer (ARG2, TA instruments, Dela-ware, USA) with cup (diameter 30 mm) and rotor (28 mm) geom-etry. The emulsionwas poured into the cup and covered with a thinlayer of low-viscosity silicone oil to prevent evaporation. Heat-setgelation was induced in situ by (1) heating the sample at a con-stant rate of 3 �C/min from 30 to 90 �C; (2) holding at 90 �C for30 min; (3) cooling at a rate of 1 �C/min from 90 to 30 �C; and (4)holding at 30 �C for 20 min. The shear storage and loss moduli weremonitored as functions of time (s). All measurements were made inthe linear viscoelastic region (0.5% strain) and at a constant fre-quency of 1 Hz. All experiments (both sample preparation andrheological measurements) were performed in triplicate.

2.4. Preparation of emulsion gels

The required quantities of solid NaCl were added to the emul-sion solutions to give final concentrations of 10, 25, 100 or 200 mM.The solutions were gently stirred to allow the NaCl to completelydissolve. The emulsions were then put in sealed plastic cylindricalcontainers (inner diameter: 25 mm) and were heated in a waterbath from 30 to 90 �C and held at 90 �C for 30 min. After heating,the gels were cooled at 4 �C. Gels containing 10, 25, 100 and200 mM NaCl were named A, B, C and D, respectively.

2.5. Measurement of mechanical properties

The cylinder samples (20 mm in height and 25 mm in diameter)were compressed between two flat plates using a TA-XT2 textureanalyser (TA instruments, Delaware, USA). Compression was

Q. Guo et al. / Food Hydrocolloids 33 (2013) 215e224 217

performed up to a strain of 50%, with the test and post-test speed of2mm s�1 and automatic trigger force of 0.1 N. The 50% compressionstrainwas chosen to avoid gel fracture. The force at the target strainwas defined as the hardness of gels. Thework to compress the gel totarget strain (Wc) and recoverable work during decompression ofthe gel (Wd) were calculated from the area under force versus timecurves. The recoverable energy (Re) was expressed as:

Re ¼ Wd=Wc (1)

The Young’s modulus was calculated from the slope of theforceetime curve in linear region.

Another large deformation test was performed using the fracturewedge set where the angle of two wedges was 30�. The dimensionsfor the samples for this test were 16�16� 25mmwidth, height andlength, respectively. The compression test was performed to a strainof 95% with the test speed of 2 mm s�1 and the trigger force was0.05 N. The fracture force and strain of gels were recorded. Thetoughness was estimated from the forceetime curve according tomethods of Agrawal et al. (1997) and Lucas and Pereira (1990). Alltestswere carriedout at roomtemperaturewith at least 10 replicates.

2.6. Determination of particle size distributions

A Malvern Mastersizer 2000 (Malvern Instruments, Worcester-shire, UK) was used to determine the average diameter and particlesize distribution of the oil droplets in the emulsion and emulsiongels. Refractive index values of 1.47 were used for the oil phase and1.33 for the dispersed (water) phase giving a relative refractiveindex of 1.105. The volume weighted mean average (D4,3) was re-ported for the oil droplet size distribution. Formeasurements of theliquid emulsion, 2 ml emulsion was added to 10 ml 5 wt% SDS so-lution and gently mixed for a few seconds before the particle sizewas measured. For measurements of the emulsion gel, 2 g gel wasadded to 20 ml solution containing SDS (5 wt %) and 2-Mercaptoethanol (2-ME) (50 mM) and shaken overnight untiltotally dispersed. This treatment dissociates the bonds in the pro-tein network and liberates the fat droplets (2-ME reduces disulfidebonds and SDS disrupts non-covalent bonds). The adsorbed pro-teins in oil-water interface are replaced by SDS. The particle sizewas measured after this treatment. For the masticated emulsiongels, SDS (about 5 wt%) and 2-ME (about 50mM)were added to themixture of bolus and debris after mastication. After the mixturewas totally dispersed, the particle size was measured. Each samplewas measured three times at room temperature.

2.7. Measurement of breakdown properties

2.7.1. Selection of panellistsThe chewing study was approved by the Massey University

Human Ethics Committee: Southern A (Application 11/60). Eightsubjects (4 males and 4 females) aged from 18 to 50 were selectedfor this study. Subject selection was based on strict dental criteria(Hutchings et al., 2011).

2.7.2. Procedures for chewing experimentsEmulsion gels were prepared and stored in individual sample

containers. Each container contained one sample of a single geltype (about 5 g). Subjects were given different containers randomlywith additional 60 ml plastic screw-cap container and 30 ml waterfor each sample. Pre-training was provided to the panellists toenable them to become familiar with the samples. Prior to chewingthe gel samples, panellists were asked to rinse their mouth withwater (3 � 30 ml). The panellists were given a sample and asked tochew the sample for as long as necessary and expectorate into the

empty container provided before they felt the impulse to swallow.Once the panellist had expectorated the sample they were asked torinse their mouth with a further 30 ml of water and the debris wascollected to another container. The chew duration and the numberof chews for each sample were recorded by a researcher.

Samples were collected from three sessions for analysing sizedistributions of fragments in masticated gel and mastication pa-rameters (4 samples for each type of gel and subject), microstruc-ture of bolus and oil droplet released from protein matrix (twosamples), and particle size distributions of oil droplets in masti-cated gels (two samples), respectively.

2.7.3. Measurement of particle size distribution of fragments withinbolus

The masticated gel including bolus and debris (two gel samples)was poured through a stack of 6 sieves of apertures 3.15, 2.00, 1.40,0.85, 0.425 and0.0380mm. Themasticated gel retained in each sievewas washed in turn for at least 2 min using mild running water andwashing bottle with gentle shaking. The fragments retained in eachsievewerewashed off the sieves and transferred to pre-weighed andpre-dried filter papers. Any particles finer than 0.0380 mm werediscardedbecause of thedifficulty of quantifying such small samples.A test using a grinder to simulate mastication showed that less than5% of the bolus weight was lost in the sieving process. The filterpapers with gel fragments were dried in an oven at 105 �C for 24 hand then weighed. The weight of dry matter retained in each sievewas expressed as a percentage of the dryweight of collected samplesin all sieves. The weight % of fragments passing each sieve was thencalculated and the mean particle size distribution of each type ofsample of every panellist was also determined (n ¼ 2).

2.7.4. Confocal laser scanning microscopyThe microstructure of boluses was studied using confocal laser

scanning microscopy (CLSM) (Leica, Heidelberg, Germany). NileRed was used to stain for oil (argon laser with an excitation line of488 nm) while Fast Green was used to stain for protein in the bolus(HeeNe laser with an excitation at 633 nm). A very small quantityof the bolus was placed on a concave confocal microscope slideabout 1.5 mm thick, mixed with 10 ml Nile Red and 10 ml Fast Green,stained for 20 min and then covered with a cover slip. The proce-dure of detecting the microstructure of emulsion gels beforemastication was the same with the exception that the sample wasallowed to stain for about 8 h.

2.7.5. Quantification of released oil dropletsThe collected bolus and debris (about 40 g as gel) were placed in

a centrifuge tube and centrifuged at 3000 g for 20 min. The processof centrifugation led to the formation of 3 separate phases; the oildroplets rose to the top, larger fragments settled to the bottom andthere was an aqueous phase in the middle. After centrifugation, thetubes were frozen (�18 �C for 24 h) which allowed the isolation ofthe top oil layer (by cutting the top). The oil phase was thawed andre-suspended in water in a glass tube (diameter: 15 mm) beforefurther centrifugation and the height of top layer was measuredafter centrifugation.

2.8. Statistical analysis

The data were analysed using PASW Statistics 18 software. One-way analyses of variance were performed to test if significant dif-ferences exist in textural parameters and mastication parameterswith gel type as factor at P < 0.05. A Pearson correlation test wasalso performed to explore the correlation between median diam-eter (d50) of particles within masticated gels and the fragmentationindex of gels (P < 0.05).

Q. Guo et al. / Food Hydrocolloids 33 (2013) 215e224218

3. Results

3.1. Emulsion gel formation

Fig. 1 shows the change in viscoelastic properties during theformation of heat-setwhey protein emulsion gels with varying NaCllevels. The varying magnitude of these results indicates the differ-ences in intermolecular bonding and microstructure. The finalstorage moduli of the whey protein emulsion gels (A, B, C and D)were 7.3�1.8, 23.0� 0.3,107.0� 5.2 and 91.0�1.1 kPa, respectively.The gelling points for gels A, B, C and D (i.e., the temperature atwhich G0 ¼ G00) were at temperatures of 85.0 � 1.3, 84.2 � 0.1,79.0 � 0.3 and 73.2 � 1.4 �C, respectively. These results suggestedthat the increase of ionic strength caused a gradual decrease in thegelling temperature of whey protein emulsions, and a marked in-crease in storage modulus up to 100 mM added NaCl. However, thefinal storage modulus of the gel containing 200 mM NaCl wassignificantly lower than that with 100mM,which is consistent withthe rheological properties of heat-set whey protein gels withdifferent NaCl concentrations reported by Ikeda et al. (1999).Confocal microscopy images (Fig. 2) revealed oil droplets (red)relatively evenly distributed in gelled emulsions AeD. These imagesalso appeared to show possible flocculation of oil droplets withinthe protein matrix. Emulsion gels AeC appeared to have a contin-uous protein network (green). By contrast, emulsion gel D did notshow a compact protein network but rather a network containingpores. The particle size distributions of oil droplets incorporated ingelled emulsionsmeasured using theMastersizer are also presentedin Fig. 2. The oil droplets within the four gels had essentially thesame particle size distribution and D4,3 (same as that of emulsions),confirming that no coalescence of droplets occurred during thepreparation of the gels.

3.2. Physical/mechanical properties of the gels

3.2.1. Large deformation properties before fractureAs shown in Table 1a, the hardness of the gel generally increased

with increasing NaCl concentration, but increasing NaCl concen-tration from 100 to 200 mM led to a significant decrease in thehardness, which was in accordance with the result of small-strainrheological measurements. The Young’s modulus showed asimilar trend. The recoverable energy decreased significantly withthe increasing NaCl concentration, indicating changes in the gel.

3.2.2. Fracture propertiesAs illustrated in Table 1b, with increased NaCl concentration, the

fracture force and toughness increased markedly. The fracture

Fig. 1. Development of viscoelastic properties of heat-set whey protein emulsion gelscontaining 10 (A), 25 (B), 100 (C) and 200 (D) mM NaCl.

strains of gels A and D were higher than those of gels B and C. Inorder to see the differences in the breakdown pattern after fracturepoint of gels with different structure, the force versus time curveswere normalized by their fracture force and fracture time (Fig. 3).The slopes of the normalized curves indicated the crack propaga-tion speed within gels (ÇakIr et al., 2012). The slopes of curves ofgels containing 100 and 200 mM NaCl were quite high, indicatingthat the gels almost totally fractured after the fracture point. Theslope of the curve of gel with 25 mM NaCl was intermediate whilethat of gel with 10 mM NaCl was low indicating the fracturepropagation of these two gels was slow after fracture.

3.3. In-mouth behaviour

3.3.1. Mastication parametersThe comparison between the mastication parameters of

different gels is presented in Table 2. The results showed thenumber of chewing cycles of gel Awas significantly lower than thatof gels B, C and D while the chew duration and chew frequency ofthe four gels were very similar with no significant differences.

3.3.2. Analysis of masticated gelExamples of boluses from gels A, B, C and D after mastication are

illustrated in Fig. 4. The boluses from gels A and B appeared to bevery wet and the fragments within them had slippery and smoothsurfaces. In contrast, the boluses of gels C and D looked very dry.Masticated gels upon collection were sieved and typical samplefragments remaining in each sieve are shown in Fig. 5. It can beseen that a larger quantity of larger fragments (i.e., larger than2.00 mm) were retained in gels A and B whereas gels C and Dretained a greater portion in sieves between 0.850 and 0.0380 mm.This result suggested that a ‘soft’ gel produced a greater quantity oflarger fragments than a ‘hard’ gel. The fragments collected wereused to determine particle size distributions, and the resultsexpressed as the percentage of dry particles mass passing eachsieve are shown in Fig. 6. The percentages of the particles largerthan 3.15 mm (not passing this sieve) for gels A and B were about63% and 47% while those for gels C and D were around 7% and 10%;By contrast, the percentages of particles smaller than 0.425 mm ofgels A and B were below 5% while those of gels C and Dwere higherthan 20%. The average data obtained from 8 subjects quantitativelysupported the qualitative data shown in Fig. 5. The particle sizedistribution curves of hard gels were near to the RosineRammledistribution whereas the shapes of curves of soft gels were nearlylinear, which is consistent with the studies of Jalabert-Malbos et al.(2007) and Hutchings et al. (2011) on natural foods. The approxi-matemedian size d50 (4.00, 2.85, 1.05 and 0.95 mm), correspondingto theoretical sieve which can pass 50% of weight, was extractedfrom curves of gels A, B, C and D. It should be noted that the totaldry sample mass recovered on sieves, expressed as a percentage ofthe dry matter ingested was quite high (90.0 � 7.5%, 86.1 � 5.5%,79.2 � 5.6% and 80.3 � 3.3% for gels A, B, C and D, respectively).

3.3.3. Oil droplet releaseConfocal microscopy images (Fig. 7) for the masticated gelled

emulsions show fragments of varying size and shape. Most of theoil droplets appeared to have been retained in the fragments of theprotein network with some released oil droplets. The particle sizedistributions of oil droplets in gels after mastication were notdifferent from those before mastication (Figs. 2 and 7), indicatingthat no coalescence happened during oral processing. Furthersupport for the low release of oil droplets from within the proteinmatrix of the masticated samples could be seen in the images ofcentrifuged boluses (data not shown), where very little oil hadcome to the surface of centrifuged samples.

Fig. 2. Confocal micrographs of heat-set whey protein emulsion gels containing 10 (A), 25 (B), 100 (C) and 200 (D) mM NaCl and particle size distributions of oil droplets within gels.Colour in red represents the oil and green the protein. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Q. Guo et al. / Food Hydrocolloids 33 (2013) 215e224 219

3.4. Correlation between physical/mechanical properties anddegree of gel fragmentation

The correlations between the degree of gel fragmentation andmechanical properties are presented in Fig. 8. The degree of gelfragmentation was represented by d50 (the higher the degree offragmentation, the lower d50). The hardness measured by the

Table 1Mechanical properties measured by the texture analyzer: (a) textural parametersmeasured by uniaxial compression test, (b) textural parameters measured by thefracture wedge set.

(a)

Gel type Hardness (N) Young’s modulus (kPa) Recoverable energy (%)

A 19.2 � 0.6* 19.7 � 2.6* 47.7 � 0.5****B 56.5 � 1.3** 60.5 � 15.1** 45.1 � 0.6***C 77.8 � 0.4**** 256.3 � 13.9**** 35.7 � 1.4**D 69.9 � 0.3*** 228.8 � 8.5*** 30.8 � 0.4*

(b)

Gel type Fracture force (N) Fracture strain (%) Toughness (J/m2)

A 3.9 � 0.3* 63.1 � 1.9** 102.9 � 2.6*B 7.0 � 0.1** 51.9 � 3.1* 146.5 � 12.6**C 12.9 � 1.4*** 55.6 � 5.0* 226.6 � 24.1***D 17.2 � 1.9**** 68.1 � 5.0*** 327.6 � 39.6****

*e**** Values with different superscripts are significantly different (P< 0.05) withinthe same groups.

uniaxial compression test did not give a significant linear correla-tion with the d50 of masticated gels (P > 0.05) (data not shown). Bycontrast, d50 had a significant negative linear correlation(r¼�0.959, P¼ 0.041) with fracture forcemeasured by the fracturewedge set (Fig. 8a).

Fig. 8b and c show that the d50 of masticated gels had a signif-icant positive linear correlation with (R/E)0.5(r ¼ 0.955, P ¼ 0.045)

Fig. 3. Normalized force versus normalized time curves after fracture point of heat-setwhey protein emulsion gels during compression testing (A: 10, B: 25, C: 100 and D:200 mM NaCl).

Table 2Mastication parameters of heat-set whey protein emulsion gels.

Gel type Number of chews Chew duration (s) Chew frequency (1/s)

A 19.4 � 2.1* 11.6 � 2.8* 1.75 � 0.33*B 24.3 � 4.3** 13.0 � 3.3* 1.93 � 0.31*C 24.8 � 4.8** 13.6 � 2.7* 1.83 � 0.20*D 23.7 � 3.8** 13.5 � 2.7* 1.78 � 0.14*

*e** Values with different superscripts are significantly different (P < 0.05) withinthe same groups.

Q. Guo et al. / Food Hydrocolloids 33 (2013) 215e224220

and a significant negative linear correlationwith (RE)0.5(r¼�0.986,P ¼ 0.014). By contrast, the d50 of masticated gels did not have avery significant linear correlation with R (P > 0.05) as shown inFig. 8d.

4. Discussion

Gel textures have been claimed to have a large influence on thenumber of chewing cycles. Foster, Woda, and Peyron (2006) havedemonstrated that both harder elastic and plastic model foodsneeded more chewing cycles. The hardness of a gel can bedescribed more accurately by linear viscoelasticity, large defor-mation properties before fracture and fracture properties together(ÇakIr et al., 2012). In the gels used in this study, hardness wasincreased with NaCl concentration and the gels containing 10/25and 100/200 mM NaCl could be characterized as soft and hard gels,respectively. As shown in Table 2, although the gel containing10mMNaCl required significantly less chewing than the others, thegels containing 25, 100 and 200 mM NaCl showed no significantdifference in the number of chewing cycles.

Liquid release may be another important factor influencing thenumber of chewing cycles (Pereira, de Wijk, Gavião, & van der Bilt,2006; van der Bilt, Engelen, Abbink, & Pereira, 2007). For gelscontaining 100 and 200 mM NaCl, large amounts of fluids werereleased from the gels during compression due to the phase sepa-ration in the gel formation compared withminimal liquid release ingels containing 10 and 25 mMNaCl. However, a significant increase

Fig. 4. Images of the boluses produced by mastication of the heat-set wh

in the number of chewing cycles was observed in the gel containing25 mM NaCl compared with that containing 10 mM NaCl andtherefore it appears that hardness rather than liquid release is thekey factor influencing the number of chewing cycles in these twogels. The fact that the number of chewing cycles does not increasein the much harder gels made with 100 and 200 mM NaCl suggeststhat liquid release may play an increasingly important role in thenumber of chewing cycles of gels. In addition, salt release mightimpact the chewing behaviours of gels containing 100 and 200 mMNaCl which have a higher salt level than saliva because sodiumrelease is negatively correlated with the number of chewing cyclesper minute (Neyraud, Prinz, & Dransfield, 2003; Pionnier et al.,2004).

The number of chewing cycles and retention time in the mouthof the four gels had no relationwith the degree of breakdown of thegels (Table 2 and Figs. 5 and 6), which implied mastication pa-rameters were not important factors influencing the fragmentationof gels during the mastication. Salty perception might also influ-ence the breakdown of heat-set whey protein emulsion gels con-taining high NaCl (100 and 200 mM) according to study of Pionnieret al. (2004). They found that high saltiness could be related to lowrate of breakdown of model cheeses. However, the gel hardnesswas generally increased with increasing NaCl concentration inpresent study, and increased hardness of heat-set whey proteinemulsion gels led to a higher degree of fragmentation, which isconsistent with the report of Peyron et al. (2004) and Jalabert-Malbos et al. (2007) on natural foods with different hardness.Especially, the fragmentation degree of four gels had a high linearcorrelation with the fracture force. As a result, gel structure is thekey factor determining the breakdown of gels in the humanmouth.That the harder gel had a higher fragmentation degree can beexplained by two mechanisms: 1, mastication strategy and 2, gelbreakdown pattern.

Stage I transport is the first step of 4 mastication sequences(stage I transport, processing, stage II transport and pharyngealswallow) which includes food intake, compressing food againstthe hard palate and tongue, and transporting food to the occlusal

ey protein emulsion gels (A: 10, B: 25, C: 100 and D: 200 mM NaCl).

Fig. 5. Photographs of the particles within each masticated heat-set whey proteinemulsion gel (A: 10, B: 25, C: 100 and D: 200 mM NaCl) retained in each sieve (0.0380,0.425, 0.850, 1.40, 2.00 and 3.15 mm, respectively) after sieving and washing.

Q. Guo et al. / Food Hydrocolloids 33 (2013) 215e224 221

surface of the molar teeth (Hiiemae & Palmer, 1999; Okada,Honma, Nomura, & Yamada, 2007). During stage I transport, thetexture of ingested food is recognized by intra-oral compressionand 10e12% strain during the tongue-palate compression wassuggested to be the critical point of gelatin and agar gels fordeciding mastication strategy for size reduction (Arai & Yamada,1993; Ishihara et al., 2012; Okada et al., 2007). Arai and Yamada(1993) used 40 subjects to observe their mastication from crush-ing to swallowing using gelatin and agar gels. They found thatthere were two types of mastication strategies and that the

strategy altered from mainly compressing with tongue and hardpalate (strategy 1) to mainly grinding with the molars (strategy 2)as the hardness and rupture stress of the model samplesincreased. The threshold of hardness for strategy alteration wasabout 1.33 � 104 Pa for gelatin gel. The hardness of the heat-setwhey protein emulsion gel containing 10 mM NaCl at the strainof 12% was about 0.5 � 104 Pa, which implied the breakdown maybe manipulated by the tongue-hard palate compression. Rupturestress is another important factor influencing the masticationstrategy. The results of the fracture wedge set showed that thefracture force of gel containing 10 mM NaCl was relatively low. Asa result this gel can be easily fragmented to smaller pieces byincisors or canines. In addition, the mastication parametersshowed there was rhythmic jaw closing and opening duringchewing. Therefore, we infer the breakdown of gel containing10 mM NaCl may be mainly manipulated by both the shearing ofincisors or canines and compression with tongue and hard palate.That the gel with 10 mM NaCl was fragmented mainly into largepieces during chewing strongly supported this inference (Figs. 5and 6). By contrast, the values of hardness of gels containing 25,100 and 200 mM NaCl at 12% strain were higher than 1.33 � 104 Paand their fracture forces, especially those of gels containing 100and 200 mM NaCl, were quite high. Therefore, these gels probablywere processed mainly by the grinding of molars thereby leadingto a high percentage of small particles during chewing (Figs. 5 and6). Moreover, the harder food needs higher chewing force andchewing activity (Foster et al., 2006; Kohyama et al., 2004; Peyron,Lassauzay, & Woda, 2002; van der Bilt et al., 2007). Therefore, thegel masticated by strategy 1 is supposed to have greater mediansize than that masticated through strategy 2.

Another important factor influencing the degree of breakdownof the whey protein emulsion gels is determined by the gelbreakdown pattern. Gels containing 10 and 25 mM NaCl had ahighly homogeneous microstructure and were soft; at 100 mMNaCl, the gel microstructure began to become hard and inhomo-geneous although this feature cannot be observed clearly inconfocal images (Pouzot et al., 2005); at 200 mM NaCl, the gel hada micro-phase separated structure with large pores (Fig. 2). Thewhey protein emulsion gels with homogeneous microstructurehad a low fragmentation degree while the gels with inhomoge-neous microstructure had a high fragmentation degree, which isagreement with the report of Gwartney et al. (2004). According tothe model of van Vliet, Luyten, and Walstra (1993), the energyapplied to deform a material (W) can be elastically stored (We),dissipated either by vicious flow of the material (Wv) or by frictionprocess between structural elements (Wc) or used to cause frac-ture (Wf):

W ¼ We þWv þWc þWf (2)

To cause fast fracture, the stress at the tip of crack should behigher than the cohesive stresses between the structural elementsand the elastically stored energy released during crack growthshould be higher than the amount of energy to cause crack growth.The storage modulus of hard gels containing 100 and 200 mM NaClwas much higher than that of soft gels containing 10 and 25 mM(Fig. 1) while the recoverable energy of hard gels (representingelastically stored energy) was a little lower than that of soft gels.Moreover, the energy consumed in growing a crack of per unit area(toughness) of hard gels was only a little higher that of soft gels(Table 1b). This suggested that during deformation the hard gel canstore high energy to cause the fast free-running crack and the crackgrowth was fast. In contrast, soft gels cannot store enough energyto cause a fast crack. In addition, the slopes of normalized forceetime curves of hard gels after fracture point were much higher than

Fig. 6. Average particle size distributions of fragments of heat-set whey protein emulsion gels (A: 10, B: 25, C: 100 and D: 200 mM NaCl) upon chewing obtained from 8 humansubjects. The points represent the amounts of material passing through a sieve of given size.

Fig. 7. Confocal micrographs of boluses of heat-set whey protein emulsion gels containing 10 (A), 25 (B), 100 (C) and 200 (D) mM NaCl together with the particle size distributions ofoil droplets within the masticated gels. Colour in red represents the oil and green the protein. (For interpretation of the references to colour in this figure legend, the reader isreferred to the web version of this article.)

Q. Guo et al. / Food Hydrocolloids 33 (2013) 215e224222

Fig. 8. Correlation between physical/mechanical properties and degree of fragmentation of heat-set whey protein emulsion gels. (a) Represents the hardness measured by thefracture wedge set; (b) represents the mechanical property index (R/E)0.5; (c) represents the mechanical property index (RE)0.5; (d) represents toughness (R).

Q. Guo et al. / Food Hydrocolloids 33 (2013) 215e224 223

those of soft gels (Fig. 3), which also suggested that hard gels breakdown fast and soft gels break slowly. The former probably causesmore small particles during mastication. Overall we suggest thatthe gel structure is the key factor determining the fragmentationdegree of heat-set whey protein emulsion gels.

According to the work of Lucas et al. (2002), there are threepatterns of crack start and growth of food particles duringchewing: 1, a crack starts and grows remote from cusps as a resultof bending against a three- (more) point cuspal support; 2, a crackis adjacent to a cusp and runs straight and rapidly through thefood particle; 3, a crack is adjacent to a cusp and the arrested crackcan only be continued by displacement of the cusp into the par-ticle. The first and third crack propagation can be represented by(R/E)0.5, which is limited by the displacement of food particles. Thesecond can be represented by (RE)0.5 which is limited by the stressimparted on the food particle. The authors also suggested differentcriteria of oral fragmentation for different foods, namely (R/E)0.5

for thick block food, (RE)0.5 for food requiring high stress to frac-ture or toughness (R) for very thin food. In our results (Fig. 8b), (R/E)0.5 had a significant positive linear correlation with the d50 ofmasticated heat-set whey protein emulsion gels, i.e., a negativelylinear correlation between degree of fragmentation and (R/E)0.5,which is in agreement with the report of Agrawal et al. (1997).This implied (R/E)0.5 is a valid criterion for estimating the frag-mentation degree of differently structured heat-set whey proteinemulsion gels during chewing. However, a better linear correlation(negative) was found between d50 and (RE)0.5, i.e., a positivelylinear correlation between degree of fragmentation and (RE)0.5.Therefore, both (R/E)0.5 and (RE)0.5 can be used to be the criteria offragmentation, which implies that the fragmentation (crackpropagation) of the four gels was complicated and the combinedmechanical property of R and E was indeed highly correlated withfragmentation of gels.

Finally, therewas a very small proportion of oil droplets releasedfrom gel matrix, which is accordance with another study (Sala, vande Velde, Cohen Stuart, & van Aken, 2007). This suggested the smalloil droplets (about 0.48 mm) within the gels are firmly bound to thegel matrix even under strong mechanical processing and the solid3-dimensional protein network surrounding the oil droplets alsoprovided a good protection against mechanical release from the gelmatrix.

5. Conclusions

This study improves the understanding the breakdown behav-iour of heat-set whey protein emulsion gels with different struc-tures in the human mouth and its relationship with the physical/mechanical properties of the gels. The gel structure which deter-mined physical/mechanical properties was the key factor influ-encing oral fragmentation of the gels. From linear viscoelasticity,large deformation properties before fracture and fracture proper-ties, the four gels can be characterized as two types: hard gel (100and 200 mM) with inhomogeneous microstructure and soft gel (10and 25 mMNaCl) with homogeneous microstructure. The hard gelshad a fast free-running crack pattern and a high degree of frag-mentationwhile soft gels had a slow crack propagation pattern anda low degree of fragmentation. The fragmentation was highly lin-early correlated with the fracture force. Furthermore, (R/E)0.5 and(RE)0.5 which are the mechanical property indices representingfood cracking in the human mouth were found to be good criteriafor estimating the degree of fragmentation of the gels. In addition,the oral processing had minimal impact on such small oil dropletswithin differently structured heat-set whey protein emulsion gels.

Acknowledgements

The authors gratefully acknowledge Jianyu Chen from theManawatu Microscopy and Imaging Centre for her help withconfocal microscopy.

References

Abhyankar, A. R., Mulvihill, D. M., & Auty, M. A. E. (2011). Combined microscopicand dynamic rheological methods for studying the structural breakdownproperties of whey protein gels and emulsion filled gels. Food Hydrocolloids,25(3), 275e282.

Agrawal, K. R., Lucas, P. W., Prinz, J. F., & Bruce, I. C. (1997). Mechanical properties offoods responsible for resisting food breakdown in the human mouth. Archives ofOral Biology, 42(1), 1e9.

Arai, A., & Yamada, Y. (1993). Effect of the texture of food on the masticatory pro-cess. Japanese Journal of Oral Biology, 35, 312e322.

Boutin, C., Giroux, H. J., Paquin, P., & Britten, M. (2007). Characterization and acid-induced gelation of butter oil emulsions produced from heated whey proteindispersions. International Dairy Journal, 17(6), 696e703.

ÇakIr, E., Daubert, C. R., Drake, M. A., Vinyard, C. J., Essick, G., & Foegeding, E. A.(2012). The effect of microstructure on the sensory perception and textural

Q. Guo et al. / Food Hydrocolloids 33 (2013) 215e224224

characteristics of whey protein/k-carrageenan mixed gels. Food Hydrocolloids,26(1), 33e43.

Chantrapornchai, W., & McClements, D. J. (2002). Influence of NaCl on opticalproperties, large-strain rheology and water holding capacity of heat-inducedwhey protein isolate gels. Food Hydrocolloids, 16(5), 467e476.

Chen, J. (2009). Food oral processingeA review. Food Hydrocolloids, 23(1), 1e25.Chen, J., & Dickinson, E. (1998). Viscoelastic properties of heat-set whey protein

isolate gels. Journal of Texture Studies, 29(3), 285e304.Chen, J., & Dickinson, E. (1999). Effect of surface character of filler particles on

rheology of heat-set whey protein emulsion gels. Colloids and Surfaces B: Bio-interfaces, 12(3e6), 373e381.

Dickinson, E., & Chen, J. (1999). Heat-set whey protein emulsion gels: role of activeand inactive filler particles. Journal of Dispersion Science and Technology, 20(1e2), 197e213.

Dickinson, E., & Hong, S. T. (1997). Influence of an anionic surfactant on the rheologyof heat-set b-lactoglobulin-stabilized emulsion gels. Colloids and Surfaces A:Physicochemical and Engineering Aspects, 127(1e3), 1e10.

Dickinson, E., & Yamamoto, Y. (1996). Rheology of milk protein gels and protein-stabilized emulsion gels cross-linked with transglutaminase. Journal of Agri-cultural and Food Chemistry, 44(6), 1371e1377.

Foegeding, E. A., Daubert, C. R., Drake, M. A., Essick, G., Trulsson, M., Vinyard, C. J.,et al. (2011). A comprehensive approach to understanding textural properties ofsemi- and soft-solid foods. Journal of Texture Studies, 42(2), 103e129.

Foster, K. D., Woda, A., & Peyron, M. A. (2006). Effect of texture of plastic and elasticmodel foods on the parameters of mastication. Journal of Neurophysiology,95(6), 3469e3479.

Gwartney, E. A., Larick, D. K., & Foegeding, E. A. (2004). Sensory texture and me-chanical properties of stranded and particulate whey protein emulsion gels.Journal of Food Science, 69(9), S333eS339.

Hiiemae, K. M., & Palmer, J. B. (1999). Food transport and bolus formation duringcomplete feeding sequences on foods of different initial consistency. Dysphagia,14(1), 31e42.

Hutchings, S. C., Foster, K. D., Bronlund, J. E., Lentle, R. G., Jones, J. R., &Morgenstern, M. P. (2011). Mastication of heterogeneous foods: peanuts insidetwo different food matrices. Food Quality and Preference, 22(4), 332e339.

Ikeda, S., Foegeding, E. A., & Hagiwara, T. (1999). Rheological study on the fractalnature of the protein gel structure. Langmuir, 15(25), 8584e8589.

Ikeda, S., & Morris, V. J. (2002). Fine-stranded and particulate aggregates of heat-denatured whey proteins visualized by atomic force microscopy. Bio-macromolecules, 3(2), 382e389.

Ishihara, S., Nakao, S., Nakauma, M., Funami, T., Hori, K., Ono, T., et al. (2012).Compression test of food gels on artificial tongue and its comparison withhuman test. Journal of Texture Studies. http://dx.doi.org/10.1111/jtxs.12002.

Jalabert-Malbos, M. L., Mishellany-Dutour, A., Woda, A., & Peyron, M. A. (2007).Particle size distribution in the food bolus after mastication of natural foods.Food Quality and Preference, 18(5), 803e812.

Kohyama, K., Hatakeyama, E., Sasaki, T., Dan, H., Azuma, T., & Karita, K. (2004). Ef-fects of sample hardness on human chewing force: a model study using siliconerubber. Archives of Oral Biology, 49(10), 805e816.

Lucas, P. W., & Pereira, B. (1990). Estimation of the fracture toughness of leaves.Functional Ecology, 4(6), 819e822.

Lucas, P. W., Prinz, J. F., Agrawal, K. R., & Bruce, I. C. (2002). Food physics and oralphysiology. Food Quality and Preference, 13(4), 203e213.

McClements, D. J., Monahan, F. J., & Kinsella, J. E. (1993). Effect of emulsion droplets onthe rheologyofwheyprotein isolate gels. Journal of Texture Studies, 24(4), 411e422.

Mehalebi, S., Nicolai, T., & Durand, D. (2008). The influence of electrostatic inter-action on the structure and the shear modulus of heat-set globular protein gels.Soft Matter, 4(4), 893e900.

Neyraud, E., Prinz, J., & Dransfield, E. (2003). NaCl and sugar release, salivation andtaste during mastication of salted chewing gum. Physiology & Behavior, 79(4e5),731e737.

Okada, A., Honma, M., Nomura, S., & Yamada, Y. (2007). Oral behavior from foodintake until terminal swallow. Physiology & Behavior, 90(1), 172e179.

Pereira, L. J., de Wijk, R. A., Gavião, M. B. D., & van der Bilt, A. (2006). Effects of addedfluids on the perception of solid food. Physiology & Behavior, 88(4e5), 538e544.

Peyron, M. A., Lassauzay, C., & Woda, A. (2002). Effects of increased hardness on jawmovement and muscle activity during chewing of visco-elastic model foods.Experimental Brain Research, 142(1), 41e51.

Peyron, M. A., Mishellany, A., & Woda, A. (2004). Particle size distribution of foodboluses after mastication of six natural foods. Journal of Dental Research, 83(7),578e582.

Pionnier, E., Chabanet, C., Mioche, L., Taylor, A. J., Le Quéré, J.-L., & Salles, C. (2004). 2.In vivo nonvolatile release during eating of a model cheese: relationships withoral parameters. Journal of Agricultural and Food Chemistry, 52(3), 565e571.

Pouzot, M., Nicolai, T., Visschers, R. W., & Weijers, M. (2005). X-ray and light scat-tering study of the structure of large protein aggregates at neutral pH. FoodHydrocolloids, 19(2), 231e238.

Rosa, P., Sala, G., Van Vliet, T. O. N., & Van De Velde, F. (2006). Cold gelation of wheyprotein emulsions. Journal of Texture Studies, 37(5), 516e537.

Sala, G., de Wijk, R. A., van de Velde, F., & van Aken, G. A. (2008). Matrix propertiesaffect the sensory perception of emulsion-filled gels. Food Hydrocolloids, 22(3),353e363.

Sala, G., van de Velde, F., Cohen Stuart, M. A., & van Aken, G. A. (2007). Oil dropletrelease from emulsion-filled gels in relation to sensory perception. Food Hy-drocolloids, 21(5e6), 977e985.

Sala, G., van Vliet, T., Cohen Stuart, M. A., van Aken, G. A., & van de Velde, F. (2009).Deformation and fracture of emulsion-filled gels: effect of oil content anddeformation speed. Food Hydrocolloids, 23(5), 1381e1393.

Sala, G., van Vliet, T., Cohen Stuart, M., van de Velde, F., & van Aken, G. A. (2009).Deformation and fracture of emulsion-filled gels: effect of gelling agent con-centration and oil droplet size. Food Hydrocolloids, 23(7), 1853e1863.

Sarkar, A., Goh, K. K. T., & Singh, H. (2009). Colloidal stability and interactions ofmilk-protein-stabilized emulsions in an artificial saliva. Food Hydrocolloids,23(5), 1270e1278.

Sok Line, V. L., Remondetto, G. E., & Subirade, M. (2005). Cold gelation of b-lacto-globulin oil-in-water emulsions. Food Hydrocolloids, 19(2), 269e278.

van der Bilt, A., Engelen, L., Abbink, J., & Pereira, L. J. (2007). Effects of adding fluidsto solid foods on muscle activity and number of chewing cycles. EuropeanJournal of Oral Sciences, 115(3), 198e205.

van Vliet, T., Luyten, H., & Walstra, P. (1993). Time dependent fracture behavior offood. InE. Dickinson, & P. Walstra (Eds.). Food colloids and polymer, stability andmechanical properties, Vol. 113, (pp. 175e199). Cambridge: Royal Society ofChemistry.

Ye, A., & Taylor, S. (2009). Characterization of cold-set gels produced from heatedemulsions stabilized by whey protein. International Dairy Journal, 19(12),721e727.