Deformation and fracture of emulsion-filled gels: Effect of oil content and deformation speed

$Page 1: Deformation and fracture of emulsion-filled gels: Effect of oil content and deformation speed$
lable at ScienceDirect

Food Hydrocolloids 23 (2009) 1381–1393

Contents lists avai

Food Hydrocolloids

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

Deformation and fracture of emulsion-filled gels: Effect of oil content anddeformation speed

Guido Sala a,b,*, Ton van Vliet a,c, Martien A. Cohen Stuart d, George A. van Aken a,e,Fred van de Velde a,e

a TI Food & Nutrition (formerly known as Wageningen Centre for Food Sciences), P.O. Box 557, 6700 AN, Wageningen, The Netherlandsb Wageningen University & Research Centre, Centre for Innovative Consumer Studies, P.O. Box 17, 6700 AA, Wageningen, The Netherlandsc Wageningen University & Research Centre, Department of Agrotechnology and Food Sciences, Bomenweg 2, 6703 HD, Wageningen, The Netherlandsd Wageningen University & Research Centre, Laboratory for Physical Chemistry and Colloid Science, P.O. Box 8038, 6700 EK, Wageningen, The Netherlandse NIZO food research, Texture Department, Kernhemseweg 2, P.O. Box 20, 6710 BA Ede, The Netherlands

a r t i c l e i n f o

Article history:Received 28 March 2008Accepted 24 November 2008

Keywords:Carrageenan gelsGelatine gelsWPI gelsEmulsionsLarge deformationsStrain hardeningElasticityViscoelasticityOil contentDeformation speed

* Corresponding author. Wageningen University &Innovative Consumer Studies, P.O. Box 17, 6700 AA, WTel.: þ31 317 482 482; fax: þ31 317 483 777.

E-mail address: [email protected] (G. Sala).

0268-005X/$ – see front matter � 2008 Elsevier Ltd.doi:10.1016/j.foodhyd.2008.11.016

a b s t r a c t

The large deformation properties of gelatine, k-carrageenan and whey protein isolate (WPI) gels filledwith bound and unbound oil droplets were studied as a function of compression speed. The rheologicalproperties of the gel matrices controlled the compression speed-dependency of the gels containing oildroplets. Polymer gels (gelatine and k-carrageenan gels) showed a predominantly elastic behaviour.Their Young’s modulus was not affected by the compression speed. The increase of fracture stress andstrain observed with increasing compression speed was related to friction between the structuralelements of the gels and, for gelatine, to the unzipping of physical bonds. Particle gels (WPI gels) showeda more viscoelastic behavior. Their Young’s modulus and fracture stress increased with compressionspeed. This was attributed to the viscous flow of the matrix and friction phenomena between structuralelements of the gel. The effect of an increase in the oil volume fraction (4) on the Young’s modulus wasfor all gels according to the Van der Poel theory. In addition, oil droplets embedded in the gel matrixacted as stress concentration nuclei and increased friction. The relative impact of these two effects wasrelated to the viscoelastic properties of the gels and to droplet–matrix interaction. For polymer gels andgels with bound droplets, stress concentration phenomena played a relatively larger role. For particle gelsand gels with unbound droplets, friction phenomena were relatively more important, increasing theviscoelastic character of the gels. As a result, an increase in 4 resulted in a decrease of both fracture stressand fracture strain for polymer gels and in an increase of the fracture stress and a decrease of fracturestrain for particle gels.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

Different kinds of foods can be categorised as emulsion-filledgels, including products like set yoghurt, fresh cheese, gelatine andstarch puddings, dairy desserts and sausages. For these foodsystems, the effects of the oil volume fraction (4) and droplet–matrix interaction on small deformation properties have receivedquite some attention in literature (Chen & Dickinson, 1998, 1999;Chen, Dickinson, Langton, & Hermansson, 2000; Dickinson & Chen,1999; Dickinson, Hong, & Yamamoto, 1996; van Vliet, 1988). Theinteraction between oil droplet and gel matrix depends on the

Research Centre, Centre forageningen, The Netherlands.

All rights reserved.

surface properties of the droplet, which are determined by thenature of the emulsifying agent (Dickinson & Chen 1999). Whenbonds are formed between the emulsifying agent and the gelmatrix, the oil droplets are bound to the matrix. In this case,depending on the ratio between the modulus of the oil droplets(Gf0 ¼ 4g/d, where g¼ surface tension and d¼ average diameter of

the oil droplet) and that of the gel matrix, the oil droplets caninduce either an increase or a decrease of the modulus of the filledgels (van Vliet, 1988). When no bonds are formed between theemulsifying agent and matrix, so that the oil droplets are unbound,added droplets always decrease the modulus of the gels (vanVliet, 1988).

In a previous article on the large deformation properties ofemulsion-filled gels measured at constant uniaxial compressionspeed (Sala, van Aken, Cohen Stuart, & van de Velde, 2007) weshowed that the interactions between oil droplet and gel matrixhave mainly an effect on the Young’s modulus of the filled gels. The

mailto:[email protected]

www.sciencedirect.com/science/journal/0268005X

G. Sala et al. / Food Hydrocolloids 23 (2009) 1381–13931382

Van der Poel theory developed for small amplitude oscillatorymeasurements in shear also holds for the Young’s modulus deter-mined by large deformation measurements. Under the experimentalconditions chosen, the fracture strain decreased with increasing 4

for droplets bound to the matrix and remained constant for unbounddroplets. The fracture stress was unaffected by 4 for bound dropletsand decreased with increasing 4 for unbound droplets.

The way food products behave when they are eaten correlatesbetter with their large deformation and fracture properties thanwith their small deformation properties (van Vliet, 2002; Van Vliet,& Walstra, 1995). The sensorial firmness of gels, for instance,depends on the apparent modulus at large deformation, and on thefracture or yield stress (Van Vliet & Walstra 1995). In order toestablish the relation between mechanical properties and sensorycharacteristics, the measurement of the mechanical propertiesshould be carried out under experimental conditions mimickingoral processing as much as possible. The fracture properties ofmany food materials strongly depend on deformation speed (VanVliet, Luyten, & Walstra, 1993). Therefore, when measuring thelarge deformation properties of gels in compression experiments,a compression speed comparable to that encountered under oralconditions (10–40 mm/s) should at least be applied (Peyron, Mio-che, Renon, & Abouelkaram, 1996). However, most data reported inliterature on the large deformation and fracture properties ofemulsion-filled gels have been obtained at much lower compres-sion speeds (Bot, vanAmerongen, Groot, Hoekstra, & Agterof,1996a;Kim, Gohtani, & Yamano, 1996; Kim, Renkema, & van Vliet, 2001;Langley & Green, 1989; McClements, Monahan, & Kinsella, 1993;Mor-Rosenberg, Shoemaker, & Rosenberg, 2004; Normand, Pluck-nett, Pomfret, Ferdinando, & Norton, 2001; Rosa, Sala, van Vliet, &van de Velde, 2006; Ross-Murphy & Todd, 1983; Sala et al., 2007;Xiong & Kinsella, 1991a, 1991b; Yost & Kinsella, 1993).

At present, relatively few data are available in literature aboutthe effects of deformation speed and droplet–matrix interaction onthe large deformation properties of emulsion-filled gels. Also thestrain-hardening and viscoelastic behaviour of these systems atlarge deformation are not well described in literature. We believethat understanding these aspects is the key to link rheological andfracture properties of emulsion-filled gels to their sensory charac-teristics, which motivated us to undertake the present study.Gelatine, k-carrageenan and whey protein isolate (WPI) werechosen as gel matrices. Emulsions made of medium-chain triglyc-eride oil, and stabilised with different emulsifying agents (WPI,Tween 20, Lactoferrin and WPI aggregates) to control droplet–matrix interaction were added to these matrices.

Gelatine gels are typically elastic polymer gels of flexible,random-coil protein chains. For gelatine gels the linear region is upto strains of about 0.5 and the gels break at true fracture strains of 1–1.5. For gelatine, values of the ratio between energy dissipated asviscous flow and energy elastically stored during deformation in thelinear region (tan d) of 0.01 and lower (frequency: 1 Hz) werereported (Takahashi, Myojo, Yoshida, Yoshimura, & Hattori, 2004).k-Carrageenan gels are also elastic polymer gels of stiff, ratherstraight chains with a linear region below a strain of 0.1. These gelsbreak at true fracture strains of 0.4–0.5. For k-carrageenan gelsprepared in demineralised water tan d increased from 0.07 to 0.09(frequency: 1 Hz) when the gelling agent concentration wasincreased from 0.3 to 1.2 wt% (Bayarri, Duran, & Costell, 2004). Acid-induced, cold set WPI gels are typically viscoelastic particle gels,with a relatively small viscous component (tan d z 0.14 at 1 Hz).The linear region is up to a true strain of 0.1 (Rosa et al., 2006).

2. Large deformation and fracture behaviour of gels

To allow a more concise discussion, in this section an overview isgiven of the relevant mechanisms responsible for the large

deformation and fracture behaviour of different types of gels,including emulsion-filled gels.

2.1. Energy balance and role of defects in the large deformationbehaviour of viscoelastic materials

To better understand the deformation speed-dependency of thefracture properties of food materials, both the energy balancerelated to fracture and the role of defects present in the structureshould be taken into consideration. The energy applied to deforma material (W) can be elastically stored (W0), dissipated either byviscous flow of the material (W 00

v ) or by friction between componentsof the system (W 00

c ) or used for fracture (Wf) (Van Vliet et al., 1993):

W ¼ W 0 þW 00v þW 00

c þWf (1)

When a piece of material is deformed, the stress at the tip of cracksand around weak spots will be higher than in the rest of thestructure (Van Vliet et al., 1993). This phenomenon is called stressconcentration. For fast fracture to occur, two requirements shouldbe fulfilled. The stress at the tip of the cracks should be higher thanthe adhesive or cohesive stresses between the structural elementsand the amount of strain energy released per unit time during crackgrowth should be higher than the amount of energy required forcrack growth. The fulfillment of this latter requirement impliesa higher W if part of it is dissipated during the deformation processand/or during fracturing. Generally, the amount of energy dissi-pated depends on the deformation speed (Van Vliet et al., 1993);hence, deformation speed is crucial.

2.2. Large deformation properties of different types of gels

The main types of food gels are polymer gels build of singlemolecules, which may vary from flexible random-coil molecules torather stiff, more rod-like ones, and particle gels formed by aggre-gation of small or large protein aggregates/particles. Typically, thesetwo types of gels respond differently to changes in deformation rate(Van Vliet & Walstra 1995). Moreover, polymer gels with covalentcross-links and gels with physical cross-links also respond differ-ently. For both types of gels, the way in which they fracture dependson the gel structure and on the strength of the bonds.

Homogeneous polymer gels with stochastically distributedcovalent cross-links have no inherent defects larger than thedistance between the cross-links, and, since the cross-links arepermanent, behave purely elastically during deformation.Furthermore, due to the low permeability of these gels, the energydissipation rate due to flow of liquid through the matrix will be low.Therefore, the fracture properties of these gels are almost inde-pendent of the deformation rate.

For polymer gels with physical cross-links, such as gelatine gels,large deformation may lead to unzipping of the cross-links.Unzipping takes a certain time, and this may result in time-dependency of the fracture parameters. Therefore, for gelatine gels,both fracture stress and strain depend on deformation rate. Bothparameters initially decrease with increasing deformation speed atspeeds lower than 2 mm/s and at higher speed increase (Bot et al.,1996a; Bot, vanAmerongen, Groot, Hoekstra, & Agterof, 1996b;McEvoy, Rossmurphy, & Clark, 1985).

In particle gels, such as casein and whey protein gels, the storagemodulus and the loss modulus depend on the type of bonds con-necting the particles within one cluster, as well as those connectingthe different clusters (Mellema, Walstra, van Opheusden, & vanVliet, 2002). In addition, the flexibility and tortuosity of the strandsin and between the clusters strongly affects the Young’s modulusand the fracture strain, while the effect on fracture stress isgenerally only small.

G. Sala et al. / Food Hydrocolloids 23 (2009) 1381–1393 1383

2.3. Effect of compression speed on Young’s modulus andfracture properties

The effects of the compression speed on Young’s modulus (Ey)and fracture properties (fracture stress, sf, and fracture strain, 3f) ofgels can be explained on the basis of a set of basic physicalphenomena related to energy dissipation mechanisms taking theenergy balance discussed above and stress concentration intoaccount. Table 1 shows a summary of the mechanisms explainingthe effect of compression speed on Young’s modulus and fractureproperties. These mechanisms are shortly described below. For anextensive description we refer to the cited literature.

For gels behaving elastically, the energy supplied during defor-mation is stored in the gel network and provides the stressesrequired to regain the original shape after removing the deforma-tion. No energy dissipation occurs and the Young’s modulus doesnot depend on compression speed. In viscoelastic gels, energydissipation occurs already in the linear region, causing the Young’smodulus to increase with compression speed. These gels exhibita more viscous behaviour, implying relatively more energy dissi-pation at low deformation speed (mechanism A1) (van Vliet, 1999).In absolute terms energy dissipation is normally higher at higherdeformation speed. It leads to lower fracture strain and higherfracture stress at higher deformation speed. With regard to thestudied gels, the energy dissipation for viscous flow will be morerelevant for particle gels as compared to the polymer gels.

Upon deformation viscous flow additional to that described bymechanism A1 may occur as a result of local yielding at placeswhere the stress is concentrated (e.g. around crack tips), leading tolocal energy dissipation (mechanism A2) (Luyten, Ramaker, & Vliet,1992; Van Vliet et al., 1993; Van Vliet et al., 1995). The global effectof this type of dissipation on large deformation properties is thesame as for mechanism A1.

For polymer gels with physical cross-links (e.g. gelatine gels),unzipping of physical cross-links and formation of new bonds canbe regarded as a kind of viscous flow as a result of (local) yieldingunder stress. These mechanisms give somewhat higher fracturestrain and fracture stress at higher deformation speed (mechanismA3) (McEvoy et al., 1985).

Energy dissipation also occurs due to friction processes betweenstructural elements of the gel network (Luyten et al., 1992; VanVliet et al., 1993; Van Vliet et al., 1995). It will be relatively higher athigher deformation speeds, compared with the other terms inequation (1). Because crack growth takes time, this gives higherfracture strain and higher fracture stress at higher deformationspeed (mechanism B).

The presence of structural inhomogeneities and defects in thegel network induces stress concentration upon deformation(mechanism C) (Luyten et al., 1992; Van Vliet et al., 1993; Van Vlietet al., 1995). This occurs at the tips of cracks or at the surface of tinyholes. It results in a lower fracture stress and fracture strain, andwill be more important in particle gels which contain relatively

Table 1Mechanisms related to the compression speed and affecting the large deformationbehaviour in the gels studied.

Affected parameter

sf 3f Ey

A1: viscoelastic behaviour þ � þþA2: induced viscous flow of the matrix þ �A3: unzipping of physical bonds* þ þB: friction between structural elements þþ þþC: stress concentration � �

sf: Fracture stress; 3f: fracture strain; Ey: Young’s modulus; þ denotes an increase ofthe parameter with speed; � indicates a decrease of the parameter. More þ or L

correspond to a stronger effect. *: Relevant for gelatin at low compression speed.

more structural defects. For viscoelastic material, deformationcauses blunting of the crack tips until the stress required for crackpropagation is reached. At higher deformation speed the bluntingof the crack tips will be less, leading to (slightly) faster crackgrowth.

2.4. Effect of the volume fraction of the droplets on Young’s modulusand fracture properties

The presence of the oil droplets in the gel matrix affects all thephenomena described in the previous section. Moreover, the oildroplets will affect the gel modulus, in a way which will depend onthe interaction between the oil droplets and the gel matrix. Theeffect of the oil droplets on the Young’s modulus can be explainedby Van der Poel theory plus extensions (mechanism D) (Smith,1974, 1975; van der Poel, 1958; van Vliet, 1988). Table 2 showsa summary of the mechanisms explaining the effect of 4 on Young’smodulus and fracture properties.

Oil droplets embedded in a gel network represent structuraldefects that induce stress concentration upon deformation(mechanism C) (Luyten et al., 1992; Van Vliet et al., 1993; Van Vlietet al., 1995). Hence, an increase in 4 represents an increase in thenumber of structural defects. At small deformation and for spher-ical holes the effective stress will increase by a factor 3 with respectto the stress in absence of structural defects. For large deformationsthe situation becomes more complex. For droplets less deformablethan the gel matrix and bound to the latter, the region between thedroplets will be deformed more than the gel matrix at the sides ofthe droplets. This gives stress concentration between the dropletsand the gel will start to fracture there as a result of de-bonding ofthe gel matrix from the droplets, or due to fracture of the gel matrixitself. This will result in both lower fracture strain and lower frac-ture stress with increasing 4 for bound droplets (mechanism Cb).For unbound droplets, fracture is likely to occur at the droplet gelmatrix interface. Moreover, for unbound droplets and especially atlow deformation speeds the extent of stress concentration betweenthe droplets will be smaller since the gel matrix can move morefreely with respect to the hard droplets (nevertheless the dropletsare not completely free due to friction effects). For unbounddroplets this will result in lower fracture strain and lower fracturestress with increasing 4, but the effect will be smaller than forbound droplets (mechanism Cu). The deformation speed effect willgive lower fracture strain and fracture stress at higher speeds.

If droplets are present in a gel matrix, viscous flow (mechanismA2) is likely to locally be more intense (Luyten et al., 1992; Van Vlietet al., 1993; Van Vliet et al., 1995). Therefore, the energy dissipationdue to viscous flow and the related dependency on compressionspeed will increase. Boundary layers between droplets and matrixcan be relatively more viscous (van Vliet, 1988); although thevolume involved is small, it may still have an effect on stress.

Table 2Mechanisms affecting the large deformation properties of emulsion-filled gelsrelated to the presence of oil droplets in the gel matrix.

Affected parameter

Bound droplets (b) Unbound droplets (u)

sf 3f Ey sf 3f Ey

A2: induced viscous flow þ � þ L

B: friction filler/matrix þþ DD

C: stress concentration LLL LLL LL LL

D: van der Poel DDD* LLL

sf: Fracture stress; 3f: fracture strain; Ey: Young’s modulus; þ denotes an increase ofthe parameter with 4; � indicates a decrease of the parameter. More þ or L

correspond to a stronger effect. *: Assuming Efiller [ Ematrix.

Table 3Volume–surface average diameter (Sauter diameter) and pH of the emulsions usedfor the preparation of the filled gels.

Emulsifier in the water phase d3,2 (mm) G0 droplet (kPa) Ey droplet (kPa) pH

1 wt% WPI 1.15 69 207 7.062 wt% Tween 1.00 20 60 4.302 wt% Lactoferrin 1.85 43 129 5.103 wt% WPI aggr. 1.45 55 165 6.64

The G0 of the droplet was calculated by G0 ¼ 4g/d whereby the surface tension g wastaken to be 20 mN/m for protein stabilised droplets and 5 mN/m for Tween stabi-lised droplets. Ey¼ 3G0 .


The presence of oil droplets embedded in the gel matrix alsoaffects the friction phenomena occurring upon deformation(mechanism B) (Luyten et al., 1992; Van Vliet et al., 1993; Van Vlietet al., 1995). As a result of inhomogeneous deformation of the gel,causing displacement of the gel matrix with respect to the dropletsurfaces, the energy dissipated by friction will increase in filled gelsas compared to gels without oil droplets.

3. Materials and methods

3.1. Materials

Porcine skin gelatine PBG 07 (bloom 280, isoelectric point 8–9)was kindly provided by PB gelatines (Vilvoorde, Belgium).k-Carrageenan was kindly donated by CP Kelco (Lille Skensved,Denmark). The k-carrageenan sample consisted of 93% mol k-unitsand 7% mol i-units, as determined by NMR spectrometry (van deVelde, Pereira, & Rollema, 2004). Powdered whey protein isolate(WPI, Bipro�) was obtained from Davisco International Inc. (LaSueur, MN, USA). Tween 20 (Polyoxyethylene sorbitan mono-laurate, in the text referred to as Tween) was obtained from Sigma(Sigma–Aldrich Chemie BV, Zwijndrecht, The Netherlands). Lacto-ferrin was kindly donated by DMV International, (Veghel, TheNetherlands). Medium-Chain Triglycerides (MCT, MIGLYOL 812 N)oil was purchased from Internatio BN (Mechelen, Belgium). Potas-sium chloride (p.a.) was obtained from Merck (Darmstadt,Germany). Glucono-d-lactone (GDL) was kindly donated by Purac(Gorinchem, The Netherlands). All materials were used withoutfurther purification. All solutions were prepared with demineral-ised water.

3.2. Sample preparation

3.2.1. EmulsionsWPI solutions were prepared by adding the protein to the

required amount of water. Subsequently, the solutions were gentlystirred for 2 h. Stock emulsions, consisting of 40 wt% MCT oil and60 wt% aqueous phase containing 1 wt% WPI, were prepared bypre-homogenising the ingredients using an Ultra-Turrax (Polytron,Kinematica AG, Lucerne, Switzerland). Pre-emulsions were furtherprocessed using a laboratory homogeniser (Ariete, Model NS1001L2 K – Panda 2 K, Niro Soavi S.p.A, Parma, Italy). The same procedurewas used for the preparation of the emulsions stabilised withTween and lactoferrin, only the emulsifying agent concentration inthe aqueous phase was 2 wt%. WPI-stabilised emulsions were usedfor the preparation of filled gelatine and k-carrageenan gels.Tween-stabilised emulsions were used for the preparation of filledgelatine gels and lactoferrin-stabilised emulsions for the prepara-tion of filled k-carrageenan gels. KCl was added to the emulsionsused for the preparation of k-carrageenan gels, up to a concentra-tion of 30 mM in the aqueous phase.

Emulsions stabilised with WPI aggregates were prepared asdescribed above, but using a 3 wt% WPI aggregates dispersion ascontinuous phase. This dispersion was prepared by heating a 9 wt%WPI solution at 68.5 �C for 2 h and subsequent cooling to roomtemperature with tap water and diluting to 3 wt%. WPI aggregates-stabilised emulsions were used for the preparation of filled WPI gels.

The droplet size distribution of the obtained emulsions wasmeasured using a Malvern Mastersizer 2000 (Malvern InstrumentsLtd., Malvern, UK). The droplet volume–surface average or Sauterdiameter (d3,2) and other characteristics of the emulsions used forthe preparation of the filled gels are reported in Table 3.

3.2.2. GelsGelatine (4 wt%) and WPI (3 wt%) gels were prepared in dem-

ineralised water. k-Carrageenan (0.6 wt%) gels were prepared in

a 30 mM KCl solution. Samples were prepared without emulsionand with different 4 (0.05, 0.11, 0.21 corresponding to oil concen-trations of 5, 10 and 20 wt%). In all samples the concentration of thegelling agent in the aqueous phase was kept constant.

For k-carrageenan and gelatine gels, the gelling agent wasallowed to hydrate for 2 h under gentle stirring at room tempera-ture. The samples were subsequently dissolved by heating at 80 �Cfor 30 min and cooled to 45 �C. In the case of k-carrageenan gels, theemulsion was heated to 45 �C prior to addition to the gelling agentsolution. For gelatine gels, the gelling agent solution was allowed tocool to 20 �C prior to mixing with the emulsion. This procedure wasfollowed to prevent depletion flocculation of the emulsion dropletsbefore gel formation (Sala et al., 2007). After mixing, the sampleswere allowed to gel at room temperatures in 60 ml plastic syringes(internal diameter 26.4 mm) coated with a thin film of paraffin oil.Gelatine gels containing Tween-stabilised emulsions were allowedto gel in a refrigerator (7 �C) for 6 h before analysis.

Gelation of the WPI gels was induced by addition of GDL(0.22 wt% in the case of WPI aggregates dispersion with a concen-tration of 3 wt%) to the WPI dispersion and to the WPI dispersion/emulsion mix and incubation at 25 �C for 17 h. The WPI dispersionwas prepared as described above. The final pH of the gels wasabout 4.8.

3.3. Large deformation experiments

Uni-axial compression tests were performed on gel pieces of25 mm height using an Instron universal testing machine (Model5543, Instron International Ldt., Edegem, Belgium) equipped witha plate–plate geometry. In order to prevent friction between theplates and the samples, the plates were lubricated with a thin layerof paraffin oil. Compression was exerted up to a strain of 80% atdifferent, constant deformation speeds: 0.05, 0.1, 0.5, 1, 2 and4 mm/s. The true strain (3H), i.e. the absolute deformation of thespecimen, and the true stress (st), i.e. the overall stress acting on thesample during deformation, were calculated as follows:

3H ¼Z H

H0

1H

dH ¼ ln�

HH0

�(2)

st ¼FA

(3)

where H0 is the initial height of the specimen, H the actual heightafter deformation, F the force measured during compression and Athe corresponding cross-sectional area of the specimen.

The determination of the recoverable energy (RE) was done bycompressing the samples up to a strain of 25% at deformationspeeds of 1 and 4 mm/s. The work necessary to compress thesamples up to this point (Wc) was calculated from the area belowthe stress vs. strain curve, and recorded. After reaching 25% strain,the samples were decompressed at the same speed and the work(Ws) released by the gel specimen was recorded. The results wereexpressed as


RE ¼ Ws

Wc(4)

3.4. Confocal Scanning Laser Microscopy (CSLM)

Samples for microstructural analyses were stained withRhodamine B (0.2 wt% solution; 10 mL per mL sample) to visualisethe protein phase. Nile Blue was chosen as a staining agent of the oildroplets. CSLM-images were recorded on a LEICA TCS SP ConfocalLaser Scanning Microscope (Leica Microsystems CMS GmbH.,Manheim, Germany), equipped with an inverted microscope(model Leica DM IRBE), used in the single photon mode with an Ar/Kr visible light laser. A Leica objective lens (63�/UV/1.25NA/waterimmersion/PL APO) was used. The excitation wavelength was set at568 nm for Rhodamine B and at 480 nm for Nile Blue. Digital imagefiles were acquired in tagged image file format and at 1024�1024pixel resolution.

4. Results

4.1. Effect of compression speed on Young’s modulus and fractureparameters

For all gels, variations in the compression speed affected at leastone of the measured parameters. For the polymer gels (gelatine andk-carrageenan), no effect of compression speed on the Young’smodulus was observed (Figs. 1A, 2A, 3A and 4A, diamond symbol).

2

3

4

5

6

7

8

Compression speed (mm/s)

Mo

du

lu

s (kP

a)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

Tru

e strain

0.0 1.0 2.0 3.0 4.0 5.0


0.0 1.0 2.0 3.0 4.0 5.0

A

C

ϕ

ϕ

Fig. 1. Gelatine bound: effect of compression speed and 4 on Young’s modulus (A), fracture sC: 0.21 oil). The arrow close to the symbol 4 indicates the increase of 4 in the samples.

However, a clear effect of compression speed on fracture stress(Figs. 1B, 2B, 3B and 4B, diamond symbol) and fracture strain wasfound (1C, 2C, 3C and 4C, diamond symbol). The effect of speed onfracture stress and fracture strain slightly differed between gelatineand k-carrageenan gels. For gelatine gels, a gradual and continuousincrease of both fracture parameters with compression speeds wasobserved. For k-carrageenan gels, an increase was found up toa speed of 1 mm/s. At higher speeds, the fracture parameters didnot increase further. The effect of compression speed on the largedeformation properties of gelatine gels is in agreement with theresults published in literature (Bot et al., 1996a, 1996b; Van Vlietet al., 1995).

For particle gels (WPI gels), the effect of compression speed wasdifferent from that described for the polymer gels. For WPI gels, theYoung’s modulus increased with increasing compression speed upto a speed of 1 mm/s, above which the value of the modulus did notincrease any further (Fig. 5A). With increasing compression speedalso the fracture stress increased (Fig. 5B). On the other hand, noeffect of compression speed on fracture strain was observed(Fig. 5C).

4.2. Droplet–matrix interactions and droplet aggregation inemulsion-filled gels

By varying the emulsifying agent used for emulsion preparation,the interactions between oil droplets and gel matrix could bemodulated. For the polymer gels studied, samples with bound aswell as unbound droplets could be prepared. For the particle gels

0

5

10

15

20

25

30


0.0 1.0 2.0 3.0 4.0 5.0

Fractu

re stress (kP

a)

Fractu

re stress (kP

a)

B

0

5

10

15

20

25

0.0 0.5 1.0 1.5 2.0Fracture strain

D

Speed

ϕ

ϕ

tress (B), fracture strain (C) and fracture points (D) (A: 0.0 oil; -: 0.05 oil; :: 0.11 oil;

2

3

4

5

6

7

8

Speed

A

C

B

D

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

Tru

e strain

0

5

10

15

20

25

30

Fractu

re stress (kP

a)

Fractu

re stress (kP

a)

0

5

10

15

20

25


0.0 1.0 2.0 3.0 4.0 5.0


0.0 1.0 2.0 3.0 4.0 5.0


0.0 1.0 2.0 3.0 4.0 5.0

0.0 0.5 1.0 1.5 2.0Fracture strain

Mo

du

lu

s (kP

a)

ϕ

ϕ

ϕ

ϕ

Fig. 2. Gelatine unbound: effect of compression speed and 4 on Young’s modulus (A), fracture stress (B), fracture strain (C) and fracture points (D) (A: 0.0 oil; -: 0.05 oil; :: 0.11oil; C: 0.21 oil). The arrow close to the symbol 4 indicates the increase of 4 in the samples.


studied, only samples with bound droplets could be prepared.As previously reported (Sala et al., 2007), the addition of a WPI-stabilised emulsion to gelatine gels resulted in filled gels containingbound droplets, whereas with the same gel matrix a Tween-stabilised emulsion gave filled gels containing unbound droplets.WPI-stabilised emulsion droplets in k-carrageenan gels behavedalso as unbound droplets (Sala et al., 2007). For the present work,lactoferrin-stabilised emulsions were added to k-carrageenan inorder to obtain filled gels with bound droplets. This is possiblebecause at the pH (7–8) of the filled k-carrageenan gels, lactoferrindisplays a positive net charge, while k-carrageenan is negativelycharged, so that the interactions between the oil droplets and thegel matrix will be attractive. These interactions indeed manifestedthemselves by an increase of the Young’s modulus with increasing4 in compression measurements (Fig. 3A). For k-carrageenansamples without oil droplets but containing the same amount oflactoferrin as filled gels with 4¼ 0.21, an increase in the Young’smodulus of 20% was observed. This increase was comparable to thatobserved for gels with 4¼ 0.05. Furthermore, no effect of lacto-ferrin on the fracture stress and fracture strain was observed(results not shown). Therefore, the effect of lactoferrin present inthe water phase of the added lactoferrin-stabilised emulsion on thestudied parameters could be neglected. For WPI gels, filled gelswith bound droplets were prepared by using a WPI aggregates-stabilised emulsion for sample preparation (Sala et al., 2007).

The microstructure of the gels, with the exception of that ofgelatine gels with unbound droplets and k-carrageenan gels withbound droplets, has been described previously (Sala et al., 2007).

The procedure chosen for the preparation of gelatine gels withbound droplets resulted in gels with non-aggregated oil droplets.For k-carrageenan gels containing WPI-stabilised emulsions,extensive aggregation of the oil droplets was observed, whilst theoil droplets embedded in WPI gels were homogeneously distrib-uted. Also for gelatine gels with unbound droplets, the oil dropletwere well dispersed or only slightly aggregated (Fig. 6A), while thelactoferrin-stabilised oil droplets present in k-carrageenan gelswere extensively aggregated (Fig. 6B). In k-carrageenan gels withlactoferrin-stabilised emulsions the aggregation of the oil dropletsis likely to be driven by electrostatic bridging between the proteinadsorbed at the oil–water interface and the polymer present insolution. For the chosen emulsions, the aggregation of the oildroplets in the k-carrageenan gels could not be prevented. Ina previous article (Sala et al., 2007) we showed that the aggregationof the oil droplets has mainly an effect on the Young’s modulus ofthe filled gels and does not affect their fracture properties. Theaggregation of the oil droplets represents an increase of the effec-tive 4, due to the matrix trapped within the aggregate. In case thestiffness of the aggregates is higher than the Young’s modulus ofthe gel matrix, this results in an increase of the Young’s modulus ofthe filled gels. In contrast, if the stiffness of the aggregates is lowerthan the Young’s modulus of the gel matrix, the increase in effective4 caused by oil droplet aggregation induces a decrease in theYoung’s modulus of the filled gel. The effect of aggregation on theYoung’s modulus can be assumed to be constant with increasing 4.Moreover, droplet aggregation occurred in both gels with boundand unbound droplets. Therefore, in this work the variations in the

0

2

4

6

8

10

12

14


Mo

du

lu

s (kP

a)

0

2

4

6

8

10

12

0.0 1.0 2.0 3.0 4.0 5.0Compression speed (mm/s)

0.0 1.0 2.0 3.0 4.0 5.0

BA

0

2

4

6

8

10

12

14

0.0 0.2 0.4 0.6 0.8 1.0Fracture strain

Fractu

re stress (kP

a)

Fractu

re stress (kP

a)

D

Speed

0

0.1

0.2

0.3

0.4

0.5

0.6

0.0 1.0 2.0 3.0 4.0 5.0Compression speed (mm/s)

Tru

e s

tra

in

C

Fig. 3. k-carrageenan bound: effect of compression speed and 4 on Young’s modulus (A), fracture stress (B), fracture strain (C) and fracture points (D) (A: 0.0 oil; -: 0.05 oil; ::0.11 oil; C: 0.21 oil). The arrow close to the symbol 4 indicates the increase of 4 in the samples.


large deformation properties of k-carrageenan gels caused by oildroplet aggregation were not investigated in detail. We study suchan effect in an ongoing work.

4.3. Effect of the volume fraction of droplets on Young’s modulusand fracture parameters

For all gels, the effect of 4 on the Young’s modulus depended onthe interaction between the oil droplets and gel matrix. In accor-dance with the Van der Poel theory, the Young’s modulus increasedwith increasing 4 for gels with bound droplets (Figs. 1A, 3A and 5A)and decreased for unbound droplets (Figs. 2A and 4A). The presenceof oil droplets did not affect the speed-dependency of the Young’smodulus of the gels.

Focussing on the effect of oil droplets on fracture parameters,differences were observed between the way gelatine and k-carrageenan gels respond to a variation in compression speed. Forgelatine gels with bound droplets, the effect of 4 on fracture stresswas small for compression speeds up to 1 mm/s. Above this speed,an increase of 4 resulted in a decrease of the fracture stress(Fig. 1B); this effect of 4 on fracture stress at higher compressionspeed was just the reverse of that on the Young’s modulus. Forgelatine gels with unbound droplets (Fig. 2B), a clear decrease offracture stress with increasing 4 was also observed, but it startedalready at a compression speed of 0.1 mm/s. Also for k-carrageenangels the effect of oil content was already visible at compressionspeeds above 0.05 mm/s both for bound and unbound droplets(Figs. 3B and 4B). For these gels, a decrease of the fracture stress

with increasing 4 was observed for both bound and unbounddroplets. For both gelatine and k-carrageenan gels, a decrease of thefracture strain with increasing concentration of bound droplets wasfound (Figs. 1C and 3C). The effect of 4 was larger at highercompression speeds, but already clear at lower speed. A decrease ofthe fracture strain with increasing 4 was also observed fork-carrageenan gels with unbound droplets (Fig. 4C). For gelatinegels with unbound droplets (Fig. 2C), the decrease of fracture straincaused by the presence of the oil droplets was only minor and notsystematically related to 4. For WPI gels with increasing 4, thefracture stress slightly increased (Fig. 5B) and the fracture strainsomewhat decreased (Fig. 5C).

Graphs of the fracture point (defined as the maximum in thestress–strain curve) at different compression speeds and forsamples with different 4 present the above given information in anefficient way (Figs. 1–5, panels D). These graphs clearly show howthe effect of 4 on fracture properties depends on the interactionsbetween oil droplets and gel matrix. Furthermore, they demon-strate that the effect of compression speed and 4 on fractureparameters is gel-type-dependent. For gelatine gels with bounddroplets (Fig. 1D), the fracture points of emulsion-filled gels fallabove those of the gels without oil. For gelatine gels with unbounddroplets (Fig. 2D), the fracture points of emulsion-filled gels fallbelow those of the gels without oil. For gels with unbound droplets,the slope of curves connecting the fracture points of samples withthe same 4 decreased. This corresponds to a decreasing effect ofcompression speed on fracture stress with increasing 4. Also fork-carrageenan gels, the fracture points of emulsion-filled gels with

00.0 1.0




2.0 3.0 4.0 5.0 0.0 1.0 2.0 3.0 4.0 5.0

0.0 1.0 2.0 3.0 4.0 5.0 0.0 0.2 0.4 0.6 0.8 1.0

1

2

3

4

5

6

7

8

9

10

0

1

2

3

4

5

6

7

8

9

10

Mo

du

lu

s (kP

a)

Fractu

re stress (kP

a)

0

0.1

0.2

0.3

0.4

0.5

0.6

Tru

e strain

A B

DC

0

2

4

6

8

10

12

14

Fracture strain

Fractu

re stress (kP

a)

Speed

ϕ ϕ

ϕ

ϕ

Fig. 4. k-carrageenan unbound: effect of compression speed and 4 on Young’s modulus (A), fracture stress (B), fracture strain (C) and fracture points (D) (A: 0.0 oil; -: 0.05 oil; ::0.11 oil; C: 0.21 oil). The arrow close to the symbol 4 indicates the increase of 4 in the samples.


bound droplets fall above those of the gels without oil (Fig. 3D),whereas the fracture point of gels with unbound droplets fall belowthose of the gels without oil (Fig. 4D). However, for these gels thecombined effect of both compression speed and 4 on fractureparameters was smaller than for the gelatine gels, mainly due to thesmaller effect of compression speed on fracture strain. The graphobtained for WPI gels was different from those for the polymer gels.For WPI gels, only a limited effect of compression speed wasobserved (Fig. 5D).

For gelatine gels, the overall effect of emulsion droplets onfracture stress and strain was similar to that reported for gelatinegels containing whey protein particles (Bot et al., 1996a) althoughwe studied a much larger filler concentration range and performedcompression measurements over a larger speed range. The simi-larity of our results and those of Bot et al. means that the effect of oildroplets on the fracture properties of these gels can be mimicked bywhey protein particles.

The strengthening of WPI gels with increasing concentration ofoil droplets stabilised by WPI aggregates is in accordance with theresults obtained by other authors for heat-induced whey proteingels containing whey protein stabilised emulsions (Langley et al.,1989; McClements et al., 1993; Xiong, Aguilera, & Kinsella, 1991a;Xiong et al., 1991).

4.4. Strain-dependency of the Young’s modulus

For several polymeric materials, the stress upon deformationincreases more than linearly with strain, a phenomenon known asstrain-hardening. This results in an increase of the modulus withincreasing strain. In order to better understand the effect of oil

droplets on the structural and functional properties of emulsion-filled gels, an analysis of the effect of both 4 and compression speedon the strain-dependency of the modulus of the gels was carriedout. The strain-hardening behaviour of the gels can be judged fromthe increase in the slope of the stress vs. strain curves withincreasing compressive strain (Fig. 7). The behaviour of the threegels was remarkably different. Both the polymer gels studiedshowed strain-hardening. Differences between the two gels can betraced to the different structure of their gel network. For gelatine,two main parts could be recognised in the stress vs. strain curve(Fig. 7): (i) at low strain (up to about 0.6), the effect of the strain onthe modulus (the slope of the stress vs. strain curve) was relativelylimited; (ii) at higher strains, the slope of the curve increased andflattened on approaching the fracture point. For k-carrageenan gels,a sudden increase of the slope of the stress vs. strain curve wasobserved at strains between 0.1 and 0.2 (Fig. 7); at higher strain theslope of the curve did not change. In k-carrageenan gels the poly-mer chains are less flexible, less random-coil-like, more straightand stiff than in gelatine gels, resulting in a fast increase of themodulus with increasing strain. This can explain the differenceobserved between the two gels. On the other hand, the modulus ofWPI gels did not change with increasing strain (Fig. 7): these gelsare non-strain-hardening.

The effect of 4 on strain-hardening behaviour was studied bysuperimposing the stress vs. strain curves of gels with different 4sand observing possible deviations in the shape of the curves (Fig. 8).To this end, the stress was normalised by dividing by the corre-sponding Young’s modulus. Fig. 8 shows the results for k-carrageenangels (the results for gelatine gels are not shown). For both polymergels studied, the initial slope of the normalised stress vs. strain curve

0

1

2

3

4

5

6

7

8

9

10


Mo

du

lu

s (kP

a)

0

1

2

3

4

5

6

7

8

9

10

0.0 1.0 2.0 3.0 4.0 5.0 0.0 1.0 2.0 3.0 4.0 5.0Compression speed (mm/s)

Fractu

re stress (kP

a)

0

0.2

0.4

0.6

0.8

1

1.2


Tru

e strain

0

5

10

15

20

25

0.0 1.0 2.0 3.0 4.0 5.0 0.0 0.5 1.0 1.5 2.0Fracture strain

Fractu

re stress (kP

a)

Speed

A B

C D

Fig. 5. WPI bound: effect of compression speed and 4 on Young’s modulus (A), fracture stress (B), fracture strain (C) and fracture point (D) (A: 0.0 oil; -: 0.05 oil; :: 0.11 oil; C:0.21 oil). The arrow close to the symbol 4 indicates the increase of 4 in the samples.


was not affected by 4; deviations with increasing 4 only emerged asthe fracture point was approached (Fig. 8A). This effect was notaffected by the compression speed (compare Fig. 8A and B) and moreevident for gels with bound droplets (compare A and C). It thusappeared that the observed decrease of the slope of the normalisedstress vs. strain curve does not show an effect of the oil droplets onstrain-hardening. As the fracture point is approached, small cracksoriginating at the surface of the oil droplets will propagate and join

Fig. 6. CLSM images of a gelatine gel containing a Tween 20-stabilised emulsion (A; staining(B; staining agent: Rhodamine B) (4: 0.21; image size 159� 159 mm).

within the gel structure, resulting in an overall decrease of themeasured stress. This effect will increase with increasing 4, inagreement with the observed effect of 4 on fracture stress and frac-ture strain. For WPI gels, which did not show strain-hardening, noeffect of 4 and compression speed was observed (results not shown).In conclusion, under the chosen experimental conditions 4 andcompression speed do not affect the strain-hardening behaviour ofthe studied gels.

agent: Nile blue) and a k-carrageenan gel containing a lactoferrin-stabilised emulsion

0

2

4

6

8

10

12

14

16

18

20

0 0.5 1 1.5 2True strain

Tru

e stress (kP

a)

Fig. 7. Stress vs. strain curves of gelatine (A), WPI (-) and k-carrageenan gels (:)(compression speed: 4 mm/s).


4.5. Recoverable energy

The energy released by a gel specimen after removing a previ-ously applied deformation gives an indication of the viscoelasticbehaviour of the gel. The lower the fraction of energy recovered, the

0

0.5

1

1.5

True strain

2

No

rm

alised

stress

0 0.2 0.4 0.6 0.8 1

0

0.5

1

1.5

2

0 0.2 0T

No

rm

alised

stress

A

C

Fig. 8. Effect of 4 on stress vs. strain curve for different types of k-carrageenan gels at difunbound droplets, 1 mm/s. (A: 0.0 oil; -: 0.05 oil; :: 0.11 oil; C: 0.21 oil).

more viscous the gels. The effect of 4 on recoverable energy (RE)differed not only between particle and polymer gels, but alsobetween the different polymer gels. For WPI gels, a gradualdecrease of the recoverable energy was observed with increasing 4

(Fig. 9). An increase in 4 resulted in less elastic and more viscousgels. For gelatine gels with bound droplets, no effect of 4 onrecoverable energy was found. However, for gelatine gels withunbound droplets, a clear decrease of the recoverable energy wasobserved with increasing 4. At higher compression speed thisdecrease was even larger (results not shown). Since 4 in gelatinegels with unbound droplets did not affect the fracture strain(Fig. 4C), it can be concluded that an increase of 4 in these gelsmade them less elastic at large deformation. For k-carrageenangels, the effect of 4 on recoverable energy was limited (Fig. 9). Forall gels, the effect of the speed was relatively small (resultsnot shown).

5. Discussion

The data currently available in literature on the large deforma-tion and fracture properties of emulsion-filled gels are ofteninconclusive and sometimes contradictory. The reason of this lackof coherency is that previous studies have usually focussed on oneparticular gel system. Moreover, large deformation and fracturemeasurements have often been performed only at one, mostly low(<1 mm/s) deformation speed. In the present study we includeddifferent gel systems, covering a representative set of cases withregard to type of gel matrices and interaction between oil dropletsand gel matrix. By measuring the effect of 4 at different compres-sion speeds, an overview could be obtained on the effect of oil

0

0.5

1

1.5

2

0 0.2 0.4 0.6 0.8 1True strain

No

rm

alised

stress

.4 0.6 0.8 1rue strain

B

ferent deformation speeds. A: bound droplets, 1 mm/s. B: bound droplets, 4 mm/s. C:

40

50

60

70

80

90

100

0 0.05 0.1 0.15 0.2 0.25

Oil volume fraction

Re

co

ve

ra

ble

e

ne

rg

y (%

)

Fig. 9. Effect of 4 on recoverable energy for different gels. A: gelatine gels with bounddroplets; >: gelatine gels with unbound droplets; :: k-carrageenan gels withbound droplets; 6: k-carrageenan gels with unbound droplets; - WPI gels with bounddroplets (compression speed: 1 mm/s).

Table 5Summary of the results with respect to the effect of 4.

Type of gel 4

sf 3f Ey RE

Gelatine bound LL LLL DDD 0Gelatine unbound LLL LL LLL LLL

k-Carrageenan bound LLL LLL DD L

k-Carrageenan unbound LL LL LL L

WPI bound DD LL þþþ LL

sf: Fracture stress; 3f: fracture strain; Ey: Young’s modulus; þ denotes an increase ofthe parameter with 4; L indicates a decrease of the parameter. More þ or L

correspond to a stronger effect; 0: no effect.


droplets on the large deformation and fracture properties of thegels (Tables 4 and 5). A key result is that measurements at highcompression speeds showed large differences between gels withregard to the effect of oil droplets, while these differences were notvisible at the low speeds usually applied in literature.

In the following we propose an explanation of the results pre-sented in this work based on the rheological properties of thesystems studied and considering the different mechanismsregarding compression speed-dependency, stress concentrationand energy dissipation upon deformation (Tables 1 and 2).

In Table 6 a summary is given of the mechanisms that mayexplain the observed effects of compression speed on the Young’smodulus and fracture properties of the gel matrices studied (Table4). For gelatine and k-carrageenan gels, the Young’s modulus wasalmost independent of the compression speed. These two gelsbehave almost purely elastically. The compression speed-depen-dency of the Young’s modulus of the viscoelastic WPI gels indicatesa more viscous behaviour (mechanism A1). These findings are inagreement with both the recoverable energy values obtained forgels without oil droplets and the tan d values reported in literaturefor these gels. For the polymer gels, the increase of fracture stressand fracture strain with increasing compression speed can beascribed to an increase of friction between the structural elementsof the gel and, for gelatine, to the partial unzipping of the physicalbond connecting flexible random-coil chains. For WPI gels, whichshowed a more viscoelastic behavior, the effect of compression

Table 4Summary of the results with respect to the effect of speed.

Type of gel Compression speed

sf 3f Ey RE

Gelatine bound DDD DDD 0 0Gelatine unbound D D 0 L

k-Carrageenan bound DDD DDD 0 0k-Carrageenan unbound D D 0 0WPI bound DD 0 DDD 0

sf: Fracture stress; 3f: fracture strain; Ey: Young’s modulus; þ denotes an increase ofthe parameter with speed; L indicates a decrease of the parameter. More þ or L

correspond to a stronger effect; 0: no effect.

speed on fracture properties can be explained on the basis of theviscous flow of the matrix and the presence of inherent defectswithin the gel network, which induce stress concentration upondeformation.

Table 7 gives an overview of the mechanisms that may explainthe effects of oil droplets on the Young’s modulus and fractureproperties (Table 5). The presence of oil droplets within the gelmatrix has an effect on the Young’s modulus, which is determinedby the stiffness of the droplet and the droplet–matrix interactions.In all gels, but especially in polymer gels, the oil droplets act asstress concentration nuclei, with differences that can be ascribed tovariations in the droplet–matrix interactions. In particle gels thepresence of bound oil droplets mainly induces an increased viscousflow of the structural elements of the gel (lower recoverable energy,Fig. 9). Moreover, in these gels the oil droplets will act as largemultiple cross-link units that fill so to speak the empty spacesbetween the aggregated protein particles. This results in a decreaseof the average size of the inherent structural defects of the gels,partly counterbalancing their effect as stress concentration nuclei.The overall effect on the fracture properties of the various gels isa combination of the described effects and of the effect of thedroplets on the Young’s modulus.

The presence of oil droplets in the gel matrix also modifies thecompression speed-dependency of the large deformation proper-ties (Table 8). Due to the presence of oil droplets, the amount ofenergy dissipated by friction increases with increasing deforma-tion speed for all gels. Nevertheless, the effect of compressionspeed on Young’s modulus and fracture parameters as describedfor the matrices does not qualitatively change as a result of thisincrease in energy dissipation, although quantitative differencescan be observed. For polymer gels (except for k-carrageenan withbound droplets), the speed effect is clearly smaller at high 4.For these gels, at least two mechanisms will affect the speed-dependency and these may compensate each other. Our data showthat up to the highest 4 studied the effect of compression speed onthe large deformation and fracture properties of the filled gels isprimarily determined by the compression speed-dependency ofthe matrix. For the polymer gels, the presence of bound dropletsdid not affect their elastic behaviour. On the other hand, thepresence of unbound droplets decreased the recoverable energy ofgelatine gels. This implies a higher amount of energy dissipated byfriction.

Table 6Summary of the mechanisms explaining the effect of the compression speed onfracture properties and Young’s modulus of the gels. For explanation of the indicatedmechanisms A1–C, see Table 1; for an extensive description, see text, Section 2.3.

Type of gel sf 3f Ey

Gelatine A3, B A3, B Elastic behaviourk-Carrageenan B B Elastic behaviourWPI A2, B, C A2, B, C A1

sf: Fracture stress; 3f: fracture strain; Ey: Young’s modulus.

Table 7Summary of the mechanisms explaining the effect of 4 on fracture properties andYoung’s modulus of emulsion-filled gels. For explanation of the indicated mecha-nisms A2–D, see Tables 1 and 2; for an extensive description, see text, Sections 2.3and 2.4.


Gelatine bound Cb, Db Cb Db

Gelatine unbound Cu, Du Cu Du

k-Carrageenan bound Cb, Db Cb Db

k-Carrageenan unbound Cu, B, Du Cu, B Du

WPI bound A2, Db, (Cb) A2, Cb Db

sf: Fracture stress; 3f: fracture strain; Ey: Young’s modulus.

Table 9Summary of the mechanisms explaining the effect of 4 and compression speed onrecoverable energy of emulsion-filled gels.

Type of gel RE

Oil droplets Compression speed

Gelatine bound Linear region, small effect Linear region, small effectGelatine unbound Energy dissipation for

friction (B)Energy dissipation forfriction (B)

k-Carrageenanbound

Outside linear reg., crackformation (*)

No effect

k-Carrageenanunbound

Outside linear reg., crackformation (*)

No effect

WPI bound Outside linear reg., crackformation (*)

No effect

*: Implies energy dissipation.


In Table 9 the mechanisms explaining the effects of 4 andcompression speed on recoverable energy are summarised. Themeasurements of the recoverable energy were performed up toa strain of 0.25. For gelatine gels, this strain is still within the linearregion. For both k-carrageenan and WPI gels, a strain of 0.25 fallsoutside the linear region. For these gels, the lower recoverableenergy as compared to gelatine is caused by the formation of crackswithin the gel structure. The formation of cracks will increase withincreasing 4. The decrease of recoverable energy in gelatine gelswith unbound droplets is related to the increase of energy dissi-pation by friction between structural elements. In these gels the oildroplets are non-aggregated. Therefore, the increase of energydissipation by friction occurs throughout the gel network, resultingin a system with a more viscous (energy dissipating) character. Thelack of this effect in k-carrageenan gels with unbound droplets canbe explained by the observation that the oil droplets are present asrelatively weak aggregates. In these gels the increase of energydissipation by friction likely remains lower as the aggregates willdeform more in accordance with the matrix than the relatively hardemulsion droplets.

The strain-hardening behaviour of the polymer gels studied islikely to be related to the stretching of the polymer chains upondeformation (Groot, Bot, & Agterof, 1996). The strain-independencyof the modulus of WPI gels is in agreement with the results ofPouzot et al. (Pouzot, Nicolai, Benyahia, & Durand, 2006) for heat-set b-lactoglobulin gels with a modulus higher than 2 kPa.

The presence of oil droplets within the gel matrix and thecompression speed do not affect the strain-hardening behaviour ofthe studied gels. As we have shown, the apparent decrease of theslope of the fracture vs. strain curve with increasing 4 at higherstrains does not represent a change of the strain-hardeningbehaviour of the gels, but is related to stress concentrationphenomena induced by the droplets. This is supported by the effectof 4 on fracture stress and strain observed for polymer (strain-hardening) gels. The calculation of strain-hardening parameters(e.g. by considering the modulus at fracture normalised by themodulus in an initial part of the stress vs. strain curve (Gwartney,D.K., & Foegeding, 2004)) without analysing the stress vs. straincurves would lead to incorrect conclusions.

Table 8Summary of the mechanisms explaining the interactions of compression speed and4 on fracture properties and Young’s modulus of emulsion-filled gels. For explana-tion of the indicated mechanisms A2–C, see Table 1; for an extensive description, seetext, Sections 2.3 and 2.4.


Gelatine bound A3 A3, B Elastic behaviourGelatine unbound A3 (*), B, Cu A3 (*), B, Cu Elastic behaviourk-Carrageenan bound B B Elastic behaviourk-Carrageenan unbound B,C B,C Elastic behaviourWPI bound A2, (B) A2, (B) A1

sf: Fracture stress; 3f: fracture strain; Ey: Young’s modulus. *: Balance between B andA3 shifted to B.

6. Conclusions

For emulsion-filled gels, the compression speed-dependency ofYoung’s modulus and fracture behaviour are primarily determinedby the gel matrix. The presence of oil droplets embedded in the gelmatrix has an effect on the Young’s modulus which is determinedby the stiffness of the droplet (Laplace pressure) and the droplet–matrix interactions. The effect of the oil droplets on the fracturebehaviour of emulsion-filled gels is related to two main mecha-nisms. Oil droplets (i) act as stress concentration nuclei (morerelevant for bound droplets) and (ii) increase the energy dissipationby friction (more relevant for unbound droplets). Stress concen-tration results in a decrease of both fracture stress and fracturestrain. Energy dissipation by friction causes an increase of bothfracture stress and fracture strain. The overall effect of stressconcentration and energy dissipation by friction induced by thepresence of oil droplets is related to the rheological properties ofthe gel matrix and to the droplet–matrix interactions. For elasticpolymer gels and gels with bound droplets, stress concentration ismore important than energy dissipation by friction. For particlegels and gels with unbound droplets, friction phenomena will bemore relevant.

References

Bayarri, S., Duran, L., & Costell, E. (2004). Influence of sweeteners on the visco-elasticity of hydrocolloids gelled systems. Food Hydrocolloids, 18(4), 611–619.

Bot, A., van Amerongen, I. A., Groot, R. D., Hoekstra, N. L., & Agterof, W. G. M.(1996a). Large deformation rheology of gelatin gels. Polymer Gels and Networks,4(3), 189–227.

Bot, A., van Amerongen, I. A., Groot, R. D., Hoekstra, N. L., & Agterof, W. G. M.(1996b). Effect of deformation rate on the stress–strain curves of gelatin gels.Journal De Chimie Physique Et De Physico-Chimie Biologique, 93(5), 837–849.

Chen, J., & Dickinson, E. (1998). Viscoelastic properties of heat-set whey proteinemulsion gels. Journal of Texture Studies, 29, 285–304.

Chen, J., & Dickinson, E. (1999). Effect of surface character of filler particles onrheology of heat-set whey protein emulsion gels. Colloids and Surfaces B: Bio-interfaces, 12, 373–381.

Chen, J., Dickinson, E., Langton, M., & Hermansson, A.-M. (2000). Mechanicalproperties and microstructure of heat-set whey protein gels: effect of emulsi-fiers. Lebensmittel Wissenschaft Und Technologie, 33, 299–307.

Dickinson, E., & Chen, J. (1999). Heat-set whey protein emulsion gels: role of activeand inactive filler particles. Journal of Dispersion Science and Technology,20(1&2), 197–213.

Dickinson, E., Hong, S.-T., & Yamamoto, Y. (1996). Rheology of heat-set emulsion gelscontaining b-lactoglobulin and small-molecule surfactants. Netherlands Milkand Dairy Journal, 50, 199–207.

Groot, R. D., Bot, A., & Agterof, W. G. M. (1996). Molecular theory of strain hardeningof a polymer gel: application to gelatin. Journal Of Chemical Physics, 104(22),9202–9219.

Gwartney, E. A., D.K., L., & Foegeding, E. A. (2004). Sensory texture and mechanicalproperties of stranded and particulate protein emulsion gels. Journal of FoodScience, 69(9), s333–s339.

Kim, K.-H., Gohtani, S., & Yamano, Y. (1996). Effects of oil droplets in physical andsensory properties of o/w emulsion agar gels. Journal of Texture Studies, 27, 655–670.

Kim, K.-H., Renkema, J. M. S., & VlietVan, T. (2001). Rheological properties ofsoybean protein isolate gels containing emulsion droplets. Food Hydrocolloids,15, 295–302.


Langley, K. R., & Green, M. L. (1989). Compression strength and fractureproperties of model particulate food composites in relation to theirmicrostructure and particle-matrix interaction. Journal of Texture Studies, 20,191–207.

Luyten, H., Ramaker, M. G., & Vliet, T. V. (1992). Fracture properties of starch gels.In G. O. Phillips, D. J. Wedlock, & P. A. Williams (Eds.), Gums and stabilizers forthe food industry, Vol. 6 (pp. 101–107). Oxford: IRL Press.

McClements, D. J., Monahan, F. J., & Kinsella, J. E. (1993). Effect of emulsiondroplets on rheology of whey protein isolate gels. Journal of Texture Studies, 24,411–422.

McEvoy, H., Rossmurphy, S. B., & Clark, A. H. (1985). Large deformation and ultimateproperties of bio-polymer gels .1. Single bio-polymer component systems.Polymer, 26(10), 1483–1492.

Mellema, M., Walstra, P., van Opheusden, J. H. J., & van Vliet, T. (2002). Effects ofstructural rearrangements on the rheology of rennet-induced casein particlegels. Advances in Colloid and Interface Science, 98(1), 25–50.

Mor-Rosenberg, Y., Shoemaker, C. F., & Rosenberg, M. (2004). Mechanical propertiesof composite gels consisting of fractionated whey proteins and fractionatedmilk fat. Food Hydrocolloids, 18(1), 153–166.

Normand, V., Plucknett, K. P., Pomfret, S. J., Ferdinando, D., & Norton, I. T. (2001).Large deformation mechanical behavior of gelatin–maltodextrin compositegels. Journal of Applied Polymer Science, 82(1), 124–135.

Peyron, M. A., Mioche, L., Renon, P., & Abouelkaram, S. (1996). Masticatory jawmovement recordings: a new method to investigate food texture. Food Qualityand Preference, 7(3–4), 229–237.

Pouzot, M., Nicolai, T., Benyahia, L., & Durand, D. (2006). Strain hardening andfracture of heat-set fractal globular protein gels. Journal of Colloid and InterfaceScience, 293(2), 376–383.

Rosa, P., Sala, G., van Vliet, T., & van de Velde, F. (2006). Cold gelation of wheyprotein emulsions. Journal of Texture Studies, 35, 516–537.

Ross-Murphy, S. B., & Todd, S. (1983). Ultimate tensile measurement of filled gelatingels. Polymer, 24, 481–486.

Sala, G., van Aken, G. A., Cohen Stuart, M. A., & van de Velde, F. (2007). Effect ofdroplet–matrix interactions on large deformation properties of emulsion-filledgels. Journal of Texture Studies, 38(4), 511–535.

Smith, J. C. (1974). Correction and extension of the Van der Poel method forcalculating the shear modulus of particulate composite. Journal of Research ofthe National Bureau of Standard, 78A(3), 355–361.

Smith, J. C. (1975). Simplification of the Van der Poel’s formula for the shear modulus ofparticulate composite. Journal of Research of theNational Bureau of Standard 419–423.

Takahashi, K., Myojo, Y., Yoshida, T., Yoshimura, K., & Hattori, M. (2004). Soft and Ca-retaining gelatin prepared by conjugating with acidic saccharides. Journal ofFood Science, 69(6), C443–C449.

van de Velde, F., Pereira, L., & Rollema, H. S. (2004). The revised chemical NMR shiftdata of carrageenans. Carbohydrate Research, 339(13), 2309–2313.

van der Poel, C. (1958). On the rheology of concentrated dispersion. Rheologica Acta,1, 198–205.

van Vliet, T. (1988). Rheological properties of filled gels. Influence of filler matrixinteraction. Colloid & Polymer Science, 266, 518–524.

van Vliet, T. (1999). Factors determining small-deformation behaviour of gels. InE. Dickinson, & J. M. Rodriguez Patino (Eds.), Food emulsions and foams: Interfaces,interactions and stability (pp. 307–317). Cambridge: Royal Society of Chemistry.

van Vliet, T. (2002). On the relation between texture perception and fundamentalmechanical parameters for liquid and time dependent solids. Food Quality andPreference, 13, 227–236.

Van Vliet, T., Luyten, H., & Walstra, P. (1993). Time dependent fracture behaviour of food.In E.Dickinson, & P. Walstra (Eds.), Food colloids and polymer, stabilityand mechanicalproperties, Vol. 113 (pp. 175–190). Cambridge: Royal Society of Chemistry.

Van Vliet, T., & Walstra, P. (1995). Large deformation and fracture behaviour of gels.Faraday Discussions, 101, 359–370.

Xiong, Y. L., Aguilera, J. M., & Kinsella, J. E. (1991). Emulsified milkfat effects onrheology of acid-induced milk gels. Journal of Food Science, 56(4), 920–925.

Xiong, Y. L., & Kinsella, J. E. (1991a). Influence of fat globule-membrane compositionand fat type on the rheological properties of milk based composite gels .2.Results. Milchwissenschaft-Milk Science International, 46(4), 207–212.

Xiong, Y. L., & Kinsella, J. E. (1991b). Influence of fat globule-membrane compositionand fat type on the rheological properties of milk based composite gels .1.Methodology. Milchwissenschaft-Milk Science International, 46(3), 150–152.

Yost, R. A., & Kinsella, J. E. (1993). Properties of acidic whey-protein gels containingemulsified butterfat. Journal of Food Science, 58(1), 158–163.

Documents

Deformation and fracture of emulsion-filled gels: Effect of oil content and deformation speed