volume 17 · issue 3 31412-1 - rheology · zusammenfassung: in der vorliegenden studie wird eine...

1 INTRODUCTIONPrevious studies have shown that using phaseseparated maltodextrin/agarose mixed gels tostudy the effect of phase volume and droplet sizeon the mechanical properties of food compositeswere made difficult because of problems withfractionation of maltodextrin [1]. It also provedimpossible to control the inclusion size indepen-

dently of the phase volume. Another approachto understand the effect of the inclusion of oneingredient in a matrix of another is to preparefilled gels, which consist of a gel matrix withinclusions of pre-made microgel particles [2]. Afew studies have been conducted using filled gelsto model food composites [3 - 6]. Ring and Stains-by [3] used different fillers to show that the small

31412-1Applied RheologyVolume 17 · Issue 3

Mechanical and Structural Properties of Maltodextrin/AgaroseMicrogels Composites

Chrystel Loret1*, William J. Frith

2and Peter J. Fryer

3

1 Unilever Foods and Health Research Institute, PO box 114, 3130 AW Vlaardingen, The Netherlands2 Unilever Corporate research, Colworth House, Sharnbrook, Bedford MK40 1DG, U.K.

3 Centre for Formulation Engineering, Chemical Engineering, University of Birmingham, Edgbaston,Birmingham B15 2TT, U.K.

*E-mail: [email protected]: +31.10.4606747

Received: 30.11.2005, Final version: 30.10.2006

© Appl. Rheol. 17 (2007) 31412-1 – 31412-13

Abstract:We present results from a new approach to the study of multicomponent gels, which allows independent inves-tigation of the effect of phase volume and droplet size of the dispersed phase on the mechanical properties ofthe mixed gel composites. The method involves preparation of agarose microgels with different sizes, which arethen embedded in maltodextrin gel matrices with different gel strengths. The effects of both phase volume anddroplet size on composite properties are dependent on the phase modulus ratio. The higher the phase modulusratio, the larger is the reinforcement effect and the effect of droplet size on mechanical properties of the mal-todextrin/agarose composites. The observed behaviour was compared with literature models for the behaviourof composite materials.

Zusammenfassung:In der vorliegenden Studie wird eine neue Methode zur Untersuchung von phasenseparierten Mehrkompo-nentengelen vorgestellt, die es erlaubt unabhängig voneinander den Einfluss von Phasenvolumen und Trop-fengröße der eingeschlossenen Phase auf die mechanischen Eigenschaften des Mehrkomponentengels zu unter-suchen. In der hier vorgestellten Methode werden zunächst Agarosemikrogele unterschiedlicher Teilchengrößeerzeugt und anschließend in Maltodextringele unterschiedlicher Gelstärke eingebettet. Der Einfluss von Pha-senvolumen und Tropfengröße auf die mechanischen Eigenschaften der erhaltenen Komposite wird dabei vomVerhältnis der Elastizitätsmoduli der einzelnen Phasen bestimmt. Je größer dieses Verhältnis ist, desto größerist der Versteiffungseffekt und der Einfluss der Tropfengröße auf die mechanischen Eigenschaften des Malto-dextrin/Agarose Gemisches. Das beobachtete Verhalten wird mit theoretischen Modellen für die resultieren-den Eigenschaften von Kompositen verglichen.

Résumé:Nous présentons des résultats issus d’une nouvelle méthode de préparation pour l’étude de gels multicompo-sés, laquelle permet de dissocier l’effet du volume et de la taille des particules de la phase dispersée sur les pro-priétés mécaniques des gels multicomposés. Des microgels d’agarose de taille différente ont été préparés puisinclus dans une matrice gélifiée de maltodextrine avec des rigidités différentes. L’effet tant du volume de phaseainsi que de la taille des particules dépend du ratio des modules des phases. L’augmentation du ratio des modulesentre les phases dispersée et continue accentue l’effet de renforcement ainsi que l’effet de la taille des parti-cules sur les propriétés mécaniques des composites de maltodextrine/agarose. Le comportement observé a étécomparé à des modèles de la littérature décrivant le comportement de matériaux composites.

Key words: biopolymer mixtures, maltodextrin, agarose, microgels, mechanical properties, microscopy

deformation properties of composites depend onthe phase volume of the filler, the deformabilityof the filler and the strength of the unfilled gelmatrix. Increasing the phase volume of the fillerresults in a reinforcement effect of the compos-ites (increase of the stiffness of the composites),which depends on the modulus of the matrix fordeformable fillers. Whilst Brownsey et al. [5] andRichardson et al. [7] did not notice a significanteffect of particle size on the dynamic moduli(they conducted studies of spherical glass parti-cles ranging from 40 to 80 mm and sephadex par-ticles ranging from 20 to 300 mm), Ross-Murphyand Todd [4], who studied glass beads rangingfrom 35 to 550 mm, found that the effect of par-ticle size depends on the volume fraction. Thisobservation was confirmed by Langley and Green[6], who studied the effect of glass spheres onwhey protein gels. At high phase volume, theeffect of particle size was enhanced and thesmaller particles increased the stiffness of thecomposites most. However, they could onlyobserve this effect if a strong interactionbetween the filler and the matrix was present.Thus, the particle size effect is not obvious anddepends on other factors such as the phase vol-ume of the filler and the interface strengthbetween the filler and the matrix. Brownsey etal. [5], Langley and Green [6] and Langely et al.[8,9] have also shown that at high deformation,the affinity between the filler particle and matrixcan profoundly affect the strength and failureproperties of the material. Most of these studiesused either a synthetic polymer such as cross-linked dextran particles or glass spheres, neitherof which are really comparable to a food fillersuch as starch granules, fat globules or seeds.Here we employ a method that allows the use offood biopolymer microgels as fillers.

We employ an emulsion route to producethe microgels. This technique has been used toprepare alginate [10], agarose and k-carrageenanmicrogels [11,12]. Frith and Norton [11] and Frith etal. [12] obtained agarose and k-carrageenanmicrogels by mixing and stirring the hot solu-tions in oil with a surfactant (admol WOL) andthen quenching the mixture. The oil and surfac-tant were then removed by washing the micro-gel suspension with water. By this method, Frithand Norton [11] were able to prepare filled gelswith two biopolymer gels involved. It was decid-ed to use the method proposed by Frith and Nor-

ton [11] to prepare maltodextrin/agarose com-posites and investigate their mechanical proper-ties. Agarose microgel suspensions wereprepared first and then incorporated in a malto-dextrin matrix. The only requirement in prepar-ing these gels was to be able to mix the two at atemperature where agarose microgel did notmelt and maltodextrin was in the solution state.

In the following, the effects of phase vol-ume, phase modulus ratio and droplet size on thelarge deformation properties of maltodex-trin/agarose composites were studied using theabove method of preparation. Results were com-pared to models for synthetic blends and com-posites. The microstructure of the compositeswas also characterised using confocal micro-scopy.

2 MATERIALS AND METHODS

2.1 MATERIALS 2.1.1 Maltodextrin and agaroseThe food grade maltodextrin (Paselli SA2) used inthis work was obtained from Avebe (The Nether-lands). Paselli SA2 is an enzymatically convertedpotato starch based product with a dextroseequivalent (DE) of less than 3. There was no glu-cose present in the sample (information from thesupplier). Paselli SA2 is available as a white pow-der, with a water content of 5 % (by dry weight).In all the figures, the term “SA2” was used as asynonym for maltodextrin. Agarose type Ia(Sigma-Aldrich, U.K.) was used. The sample hada sulphate content of less than 0.2 % and an ashcontent lower than 0.6 % (information from sup-plier). In all the figures, the term “Ag” was usedas a synonym for agarose.

2.1.2 Preparation of agarose microgelsAgarose microgels were prepared by an emulsifi-cation method that has been reported previously[11, 12]. 8% agarose solutions were prepared by dis-persing the required amount of powder in de-ionised water (18 MW/cm) at room temperaturefor 10 minutes and then heating in a boiling waterbath for 30 minutes. This solution was poured pro-gressively into sunflower oil in a 1:2 ratio. The sun-flower oil had been pre-heated at 80˚C to avoidthe gelation of agarose. To facilitate emulsifica-tion, a strong lipophilic water-in-oil emulsifier(Hydrophilic Lipophilic Balance (HLB) value of

31412-2 Applied RheologyVolume 17 · Issue 3

around 4 - 5), admul WOL (1 %) (Quest interna-tional, The Netherlands) was dissolved in the oilphase. Whilst the agarose solution was added tothe oil, the mixture was stirred with an L4R Silver-son homogeniser at different speeds to create dif-ferent agarose droplet sizes. In order to preparethe smaller sized microgels, the rotor of the Sil-verson homogeniser was fitted with the suppliedmesh screen, containing round holes of diameterof 1 mm. For the largest sized microgel, the screenwas removed. Three different samples were pre-pared: “high shear” microgels (maximum speed ofSilverson homogeniser ª 8000 rpm), “low shear”microgels (minimum speed of Silversonhomogeniser ª 2000 rpm) and “no screen” micro-gels (minimum speed without screen). The mix-tures were stirred during 30 minutes to allow asteady state droplet size to develop. The emulsionwas then left to cool in a beaker filled with ice andstirred slowly, to prevent aggregation and coales-cence. On cooling, the aqueous polysaccharidephase gelled, forming agarose microgel particles.The sample was then centrifuged to separate themicrogel from the oil phase (Sorvall centrifuge, RC3C, U.K.). The speed of the centrifugation useddepended on the size of the agarose dropletsformed. The speeds selected were those whereclear separation could be observed between theoil top phase and the sedimented microgels.Microgel suspensions were also observed underphase contrast microscopy to check if they stayedintact after the centrifugation. Table 1 summaris-es the speeds applied for the different samples.

As produced, the sedimented microgel solu-tion has got still a significant amount contami-nation of oil. Therefore, a second emulsificationprocess was conducted using 1% tween, whichfavours an oil in water emulsion. Samples werethen subsequently purified by repeated cen-trifugation (5 times for 15 minutes) and redisper-sion in de-ionised water to eliminate the surfac-tant and oil left. The speed of centrifugationapplied to the different samples is the same asmentioned in table 1.

Earlier studies [12] showed that once gelled,the microgels have little tendency to swell orshrink, thus the concentration of biopolymerinside the microgel particle should be the sameas that of the original solution, here 8%. Thisallows the volume fraction of an aqueous sus-pension of microgels to be determined from theweight solids: f = Ws/W0, where f is the volume

fraction of the microgel suspension, and Ws andW0 are the weight concentrations of the micro-gel suspension and the original polymer solution,respectively. A microgel stock was produced bydiluting the sediment in distilled water from thelast centrifugation step to a constant concentra-tion of microgel (60 %) as determined by dryweight measurement (16 h at 60˚C). Therefore,the agarose microgel suspension concentrationfor the three samples with different droplet sizeswas 4.8 % (w/w).

2.1.3 Preparation of maltodextrin/agarosemicrogel solutionsComposites of maltodextrin/agarose microgelswere prepared by mixing 51 g of the stock micro-gel with 170 g of maltodextrin solutions at dif-ferent concentrations. Agarose microgel suspen-sions were composed of 60 % by volume.Therefore when the maltodextrin solution wasmixed with the microgel suspension, the solu-tion was diluted. Hence, for initial concentrationsof 30, 35, and 40 % maltodextrin, the concentra-tion of maltodextrin in the mixed solutions wasrespectively estimated to be 25.5, 29.5, and 33.9%. The maltodextrin and agarose microgel solu-tions were mixed for 15 minutes at 60˚C, at whichtemperature the agarose microgels do not meltand maltodextrin solutions do not gel [13]. Themixture was then centrifuged at 60˚C for 45 min-utes (Beckman, L-8M, U.S.A.), using a speed thatdepended on the size of the agarose microgels(see Table 2).

The two phases (maltodextrin as super-natant and agarose microgel suspension as sed-iment) were separated, stored in the oven at


Microgel Treatment

“High shear” 5 200 g – 1h – room temperature“Low shear” 1 900 g – 1h – room temperature“No screen” 500 g – 1h – room temperature

Composites Treatment

maltodextrin/“High shear” microgel 13000 g – 45 min – 60ºCmaltodextrin/“Low shear” microgel 3300 g – 45 min – 60ºCmaltodextrin/“No screen” microgel 800 g – 45 min – 60ºC

Table 1 (above):Speed applied during cen-trifugation for differentagarose microgel solutions:“high shear” microgels, “lowshear” microgels and “noscreen” microgels.

Table 2:Speed applied during cen-trifugation for differentmaltodextrin/agarosemicrogel solutions: “highshear” microgels, “lowshear” microgels and “noscreen” microgels.

60˚C and mixed again in different proportions.This centrifugation procedure was used in orderto always have the same concentration in eachphase when they are mixed in different propor-tions. Knowing the agarose concentration of theinitial microgel samples (4.8 %) and the quanti-ties of supernatant and sediment obtained, thevolume fraction of agarose microgel in the sedi-ment could be determined as could the phasevolume of agarose microgels in each of the sam-ples prepared.

2.1.4 Preparation of maltodextrin and compos-ites gel samplesThe mixed solutions were poured into moulds(13.5mm deep and 12.3mm diameter), which wereinitially stored in the fridge. They were overfilledto avoid air bubbles and the excess gel was forcedout while sealing. The moulds were storedovernight, in a water bath at 10˚C. Confocalmicroscopic examination of the top and bottomsurface of the mixed gels as well as a sectionthought the samples confirmed a uniform distri-bution of the agarose microgels within the mal-todextrin matrix.

2.2 METHODS 2.2.1 Large deformation measurementsUniaxial compression experiments were per-formed on an Instron Universal Testing Machine,Model 4501 (Instron SFL, U.K.), equipped with anenvironmental chamber held at 10˚C. The com-pression tests were carried out with a 100 Ncapacity load cell at a displacement rate of 25mm/min. The samples were extracted from theperspex moulds after 18 h and aligned in the cen-tre of the stainless steel compression platens. Adefined distance between the sample and thetop plate was left before compression to avoidany compression before measurement. In orderto ensure slip boundary conditions, the plateswere lubricated with mineral oil. A minimum offive compression tests was performed for eachexperimental condition to maintain reasonablestatistical validity and analyse the variability inthe data.

Corrected true stress and strain were calcu-lated due to the potentially large change in sam-ple cross-sectional area at high deformations (i.e.> 0.1 or 10%) [14]. The corrected “true” stress isapproximated by:

(1)

where F is the load applied, A0 is the original spec-imen cross-sectional area, H is the instantaneousspecimen height and A0 is the original height. Thecorrected “true” (or Hencky) strain, e, is defined as:

(2)

The apparent compressive elastic modulus, Eapp(Pa), is then given by the initial linear region ofthe stress/strain curve following:

(3)

The apparent compressive elastic modulus wasmeasured within a low strain region, i.e. between0.02 to 0.05. The true stress and strain at failurewere taken as the maximum of the truestress/strain response, sf and ef, and their rela-tive values, sr and er, were calculated as follows:

(4)

(5)

Also, the relative apparent elastic modulus wasdefined as:

(6)

The subscript 0 refers to the mechanical proper-ties of the matrix phase.

2.2.2 Confocal Scanning Laser Microscopy (CSLM)Maltodextrin/agarose microgel compositeswere observed under confocal microscopy. Thesystem consisted of a MRC 600 CSLM (Bio-RadLaboratories, U.K.) attached to an Ortholux IImicroscope (Leica Microsystems, U.K.). The CSLMwas operated with an argon-krypton laser withan excitation at 568 nm and an emission collect-ed at 585 nm. A 20x objective lens with no zoom


was used. Structural observation was typicallyconducted about 100 mm underneath the top sur-face of the sample being examined. Imaging atthis depth minimised any surface related opticalartefacts. To increase the signal to noise in theimages, multiple scans were averaged (Kalmanfiltered) during data collection. The sampleswere stained with Rhodamine-B to provide a highlevel of contrast between the agarose and mal-todextrin phases. For the micrographs presentedhere, the maltodextrin stains more intensely andappears light, whereas the agarose appears dark.

2.2.3 Size distribution of agarose droplets in mal-todextrin/agarose composites using a mastersizer To determine the size distribution of the agarosemicrogel, a Mastersizer c (Malvern Instrument,UK) was used. A 300 mm lens, which allows theparticle sizes between 1.2 and 600 mm to be mea-sured, was set up and a sample cell with a 2.6 mmpath length was selected.

3 RESULTS AND DISCUSSION

3.1 EFFECT OF PHASE VOLUME ON LARGE DEFOR-MATION PROPERTIES OF MALTODEXTRIN/AGA-ROSE COMPOSITESIn this section, we discuss, as one example, theeffect of phase volume on large deformationproperties of maltodextrin/agarose compositesusing composites consisting of the “low shear”microgel in 30% maltodextrin gel. Figure 1a andb show the effect of agarose microgel phase vol-ume on the true stress/strain response of mal-todextrin/agarose composites prepared at dif-ferent phase volumes for one size of microgels(“low shear” microgel) and one matrix concen-tration (30 %) using the Instron. For clarity, dataare given on two figures. Two types of behaviourare seen: (i) at low phase volume (f < 28 %), thecomposites show a true stress/strain response

similar to the maltodextrin gel, with a strain soft-ening behaviour, i.e. the gradient of stress versusstrain after a linear response decreases beforefracture and (ii) at high phase volume (f > 28 %),a linear relation between stress and strain isobserved nearer fracture. This behaviour is simi-lar to that of agarose gels at high concentrations[15]. Increasing the phase volume of agarosemicrogels changes the true stress/strain behav-iour of the mixed composites bringing it closer tothat of the filler.

From Figure 1a and b, the apparent elasticmodulus and the strain and stress at failure canbe determined. Table 3 gives the values of the dif-ferent parameters as a function of agarosemicrogel phase volume. The relative values ofthese parameters are plotted as a function of thephase volume of agarose microgels in Figures 2,3, and 4 (EAg/ESA2 = 9). Increasing the phase vol-ume of agarose microgel increases the apparentelastic modulus and the true stress at failure anddecreases the true strain at failure. Compositeswith high phase volume of agarose microgel arethus stiffer and less deformable.

In contrast to the method of preparation ofthe composites by phase separation and remixreported in an previous article [2], this inclusionof agarose microgel in maltodextrin matrixallows the effect of phase volume on mechanical


Figure 1:True stress/strain responseof 30% maltodextrin / 8%agarose composites at dif-ferent volumes of agarosemicrogels (denoted % Ag)(displacement rate:25 mm/min, Tcabinet : 10˚C).

Table 3:Values of the apparent elas-tic modulus, the true stressand strain at failure of 30%maltodextrin / 8%agarosecomposites at differentagarose microgel phase vol-umes. The errors were deter-mined by repetition of themeasurements five times.

Agarose Apparent True stress True strainphase volume elastic modulus at failure at failure(%w/w) (kPa) (kPa)

0 224.6 ± 7 17.1 ± 1.07 0.1306 ± 0.00987 291.1 ± 8.1 21.3 ± 0.87 0.1151 ± 0.008417 373.9 ± 10 25.2 ± 1.7 0.1038 ± 0.008428 599 ± 9.6 37.5 ± 1.13 0.098 ± 0.002535 765.5 ± 10 41.5 ± 0.6 0.0902 ± 0002543 799.8 ± 10 47.1 ± 0.5 0.0892 ± 0.003853 1056.3 ± 54 73.4 ± 0.32 0.0843 ± 0.0055

a) b)

properties of the mixed composites to be stud-ied without an influence of the droplet size. How-ever as noticed in the literature, this reinforce-ment effect observed may also depend on thephase modulus ratio and droplet size. The influ-ence of these two parameters on the large defor-mation properties of maltodextrin/agarose com-posites are discussed in section 3.2 and 3.3.

3.2 EFFECT OF PHASE MODULUS RATIO ONLARGE DEFORMATION PROPERTIES OF MALTO-DEXTRIN/AGAROSE COMPOSITESTo study the effect of the phase modulus ratio(EAg/ESA2) on the large deformation properties ofmaltodextrin/agarose composites, composites

were prepared by mixing “low shear” 8 % agarosemicrogels with maltodextrin solutions at differ-ent concentrations.

3.2.1 Mechanical properties of agarose microgelsand maltodextrin matricesAgarose gels in excess water do not swell [12].Therefore, the polymer concentration within themicrogel is supposed to be the same as the bulkconcentration of the solution used to make it,here 8%. Therefore, the apparent elastic modu-lus of the agarose microgel was approximated tobe the same as 8% agarose, i.e 2060 ± 165 kPa.

Large deformation measurements were con-ducted on the gelled maltodextrin matrix (the topphase obtained after centrifugation of the initialmaltodextrin/agarose mixture). Table 4 summaris-es the apparent elastic modulus and the stress andstrain at failure for the different matrices. In a pre-vious study [1], a power law relation was establishedbetween the concentration and the apparent elas-tic modulus of maltodextrin gels: Eapp = 10-4.7c6.9.Assuming that the top phase is only composed ofmaltodextrin, the concentration of the maltodex-trin gel matrix can be determined using this equa-tion and the data from Table 4. Concentrations arevery close to the estimated concentrations (seeTable 5), which suggests both that solutions of mal-todextrin do not diffuse into the agarose particlesand that agarose microgels do not swell or shrink.From the previous data, the values of the modulusratio for the different composites prepared can becalculated and are summarised in Table 6.

3.2.2 Effect of phase modulus ratio on largedeformation properties of maltodextrin/agarose composites prepared at different phasevolumesThe relative values of apparent elastic modulus,true stress and strain at failure as a function ofphase volume for different phase modulus ratiosare plotted on figure 2, 3 and 4, respectively. Thesefigures show that increasing the modulus ratiobetween the agarose microgel and the maltodex-trin gel matrix affects the reinforcement proper-ties of agarose microgels. Composites with a smallphase modulus ratio show a smaller relativeapparent elastic modulus than those with a highphase modulus ratio. This difference is enhancedfor high phase volumes of agarose microgel. Thesame effect of the phase modulus ratio on the truestress at failure can be observed. The effect of


Figure 2:Comparison of the relativeapparent elastic modulusbehaviour of maltodex-trin/“low shear” agarosecomposites with differentphase modulus ratios andphase volumes witha) the modified Kernermodel (Eq. 8),b) the Sudduth model (Eq. 9)andc) a modified isotress model(Eq. 10).

a)

b)

c)

modulus ratio on the relative true strain at failureis less pronounced than on the two other para-meters. At agarose microgel phase volumes lowerthan 0.2, the phase modulus ratio does not seemto influence the true strain at failure of the com-posites. When the phase volume of agarose micro-gels is increased, composites with the higher mod-ulus ratio show the higher relative true strain atfailure. These results are in accordance with Ringand Stanley [3], who showed that reinforcementby a deformable filler (sepharose) becomes lesspronounced when the gel matrix (gelatin) isstronger.

3.2.3 Comparison of the experimental data withdifferent models from the literatureA very extensive literature exists on the theoryand use of particle fillers in polymeric materials[16 - 18]. Although the forces applied are muchsmaller than for plastics and metals, an engi-neering approach to model the elasticity andfracture of food model composites should be pos-sible. A variety of approaches have been pro-posed to predict the mechanical properties offilled systems, ranging from curve fitting tosophisticated analytical treatments. Evaluationof semi-empirical relationships, which rely uponthe determination of adjustable parameters bycurve fitting techniques, is beyond the intent ofthis work. Rather, attention will be directed tothose models for which the parameters can befixed by other considerations so that the result-ing relationships can be used predictively.

Models for the relative apparent elastic modulusAs the apparent elastic modulus is related to theshear modulus by the following equation:

(7)

Models developed for the dynamic modulusshould be applicable to the apparent elastic mod-ulus. One of the most versatile and elaborateequations for a composite consisting of spheri-cal particles in a matrix is due to Kerner [19]. Thismodel was improved by Nielsen [20], who includ-ed the maximum packing fraction parameter.Here, the maximum packing fraction was esti-mated to be 0.7. The modified Kerner equationwas written as follows:

(8)

where

where G and G0 are respectively the shear mod-ulus of the composite and the matrix, ff and fmare respectively the phase volume of the filler andthe matrix, n0 is the Poisson ratio.

Figure 2a shows that Eq. 8 does not fit thedata, either at low or high phase volume. Themodel first underestimates the data, then at highphase volume, it overestimates them. However,the model does capture the effect of modulusratio quite well.

Two other models were considered, the Sud-duth model [21] and the isostress model [22], withno better success (Figures 2b and c). Sudduth alsodeveloped an equation taking into account themaximum packing fraction using a generalisedviscosity model based on the Smallwood equa-tion [23]:

(9)


Initial concentration Apparent elastic True stress True strain ESA2/ESA2of maltodextrin modulus ESA2 at failure at failuresolution(%w/w) (kPa) (kPa)

30 77.5 ± 3.6 6.9 ± 3 0.1528 ± 0.0086 2735 225 ± 7 17 ± 1 0.1306 ± 0.0098 940 1000 ± 23 85 ± 4 0.1327 ± 0.0068 2

Initial concentration Concentration Concentrationof maltodextrin expected after using Eq. solution dilution Eapp = 10-4.7·c6.9

(%w/w) (%w/w) (%w/w)

30 25.5 24.235 29.5 2940 33.9 35.5

Table 4 (above):Apparent elastic modulusand fracture properties ofgelled top phases (mal-todextrin matrices) at dif-ferent concentrations, andphase modulus ratios formixture of 8 % agarosemicrogel with maltodextrinat different concentrations(EAg/ESA2), where elasticmodulus of the agarosemicrogel is approximated tobe 2030 ± 165 kPa. Theerrors were determined byrepetition of the measure-ments five times.

Table 5:Comparison of the estimat-ed concentration of mal-todextrin gel matrix due todilution of mixing with theagarose microgel suspen-sion, with the concentrationcalculated using equationrelating the apparent elasticmodulus.

where

Concerning the isostress model [22], Frith andNorton [11] modified it to describe the behaviourof the apparent elastic modulus for agarose/ge-latin and k-carrageenan/gelatin compositesreplacing the volume fraction by the reduced vol-ume fraction (f/fm) and assuming an infinitevalue for the filler modulus. In this study, the fillermodulus was not assumed to be infinite and themodified equation of the isostress was written:

(10)

Models for the fracture properties (relative stressand strain at failure)Models for fracture properties of composites,such as the true stress and strain at failure, havebeen less studied than those for the apparentelastic modulus. Nielsen [24] proposed a modelto describe the behaviour of the true strain at fail-ure as a function of spherical filler phase volumefor composites where particles and matrixadhere together. Ross-Murphy and Todd [4] pro-vided the equation for a non adhesive compos-ite. The equations are for composites with adhe-sive particles:

(11)

and for composites with no adhesive particles:

(12)

where S is the stress concentration factor intro-duced by Nielsen [24]. For simplicity, Ross-Murphyand Todd [4] assumed that Sequals 1. Figure 4 showsthe behaviour of the relative true strain at failureas a function of phase volume for maltodex-trin/agarose composites with different phase mod-ulus ratios. Equations 11 and 12 are also plotted.

Equations 11 and 12 do not include phasemodulus ratio or maximum packing fraction,therefore, they would not be expected todescribe the data well. Of the two models, Eq. 12for non adhesive particles describes best thebehaviour of the true strain at failure as a func-tion of agarose microgel phase volume. The Ross-Murphy model, assuming no adhesion betweenthe particles and the matrix, fits the data for com-posites with a high phase modulus ratio very wellindeed. This model is valid when there is a larg-er difference between the modulus of the fillerand the one of the matrix. These results suggestthat the weak interfacial strength alreadydescribed by Loret et al.[22] controls the failurebehaviour.

Concerning the true stress at failure as afunction of filler phase volume, some authors[24 - 27] suggested that, assuming no adhesionbetween the matrix and the filler and no stressconcentration effects, the relative strength couldbe described by a power law relation:

(13)

where sc and s0 are the strengths of the com-posite and the matrix, respectively, ff is the vol-ume fraction of the filler and b and n are con-stants depending on the assumed particle shapeand arrangement in the model composite. For acubic matrix with uniformly dispersed sphericalparticles, Nicolais et al. [25] transformed Eq. 13 by:

(14)


Figure 3 (left):Comparison of the relativetrue stress failure behaviourof maltodextrin / “lowshear” agarose compositeswith different phase modu-lus ratios and phase vol-umes with Nicolais, Ross-Murphy and Langley models(Eqs. 14, 17, and 18).

Figure 4:Comparison of the relativetrue strain failure behaviourof maltodextrin / “lowshear” agarose compositeswith different phase modu-lus ratios and phase vol-umes with Nielsen and Ross-Murphy models(respectively Eqs. 11 and 12).

These models predict that the true stress at failuredecreases with increasing filler phase volume.

Another approach was proposed by Ross-Murphy and Todd [4]. As the relation of the rela-tive stress at failure to the modulus and the rel-ative strain at failure is: sr = Er·er, using theNielsen model for relative strain at failure as afunction of phase volume (Eqs. 11 or 12, assumingS equals 1) and Landel equation [28] (Eq. 15) todescribe the modulus as a function of phase vol-ume, Ross-Murphy and Todd [4] developed newmodels to describe the stress at failure for adhe-sive and non adhesive composites. The Landelequation is:

(15)

Therefore, for adhesion between the filler andthe matrix, the relative stress at failure can bedefined for good interfacial adhesion by:

(16)

and for no interfacial adhesion by:

(17)

In the case of good interfacial adhesion, when ffis less than 0.42, the relative stress at failure islower than for the modified matrix, whereas forff > 0.42, it increases. Whereas, in the case of nointerfacial adhesion, the stress at failure isalways increased with ff compared to theunfilled matrix. Since it appears that there is aweak interface between the filler and matrix inthe maltodextrin/agarose composites [2], onlythe models for non adhesive particles will be con-sidered.

Langley and Green [6] modified the Ross-Muphy model [4] to take into account the maxi-mum packing fraction. They showed that data onSephadex/whey protein gel composites withweak interaction between the components werebetter fitted by:

(18)

Figure 3 shows the relationship between truestress at failure and phase volume and phasemodulus ratio, together with Nicolai, Ross-Mur-phy and Langley models (Eqs. 14, 17, and 18). Thetrue stress at failure increases monotonicallywith the phase volume of agarose microgels. TheNicolais model (Eq. 14) thus has the wrong trend.Equations 17 and 18 are also plotted to show theeffect of the introduction of the maximum pack-ing fraction parameter in the model.

As can be seen in Figure 3, the Ross-Murphymodel (Eq. 17) best describes the behaviour of thetrue stress at failure as a function of phase vol-ume of agarose microgels for composites withlow phase modulus ratio. In contrast, compositeswith high phase modulus ratio are betterdescribed by the Langley model (Eq. 18), althoughthat at high phase volume this model overesti-mates the relative true stress at failure.

3.3 EFFECT OF MICROGEL SIZE ON LARGE DEFOR-MATION PROPERTIES OF MALTODEXTRIN/AGA-ROSE COMPOSITES3.3.1 Size distributions of agarose microgelsSize distributions of the microgel suspensionsare shown in Figure 5. In this figure, two sets ofpreparation are shown indicating the repro-ducibility of the process. Figure 10 shows thatagarose microgels could be prepared with differ-ent sizes from 5 to 150 mm. In each case, the micro-gel suspensions show a broad monomodal dis-tribution. The “no screen” agarose microgelsuspension shows a broader distribution thanthe two others, and its reproducibility seemssomewhat poorer.

3.3.2 Confocal microscopic observation of themaltodextrin/agarose composites with differentagarose droplet sizesFor all samples, the composite microstructurewas observed under confocal microscopy. Exam-ples of the microstructures obtained for the


Figure 5:Size distribution of 8 %agarose microgel suspen-sions prepared using a Sil-verson homogeniser at dif-ferent speeds: low speed(“low shear” microgel), highspeed (“high shear” micro-gel) and low speed withoutscreen (“no screen” micro-gel). The mean averagediameters are given for thedifferent microgel suspen-sions. The open symbolsshow size distribution forsamples prepared on anoth-er day.

30%maltodextrin / 8%agarose composites madewith different size of agarose microgels and atdifferent phase volumes is given. The observa-tions were identical for the composites preparedat lower and higher phase modulus ratio(35%amltodextrin / 8%agarose and 40%mal-todextrin / 8%agarose).

Figure 6 shows that (i) increasing the phasevolume of the agarose microgels, increases thenumber of droplets present in the samples and(ii) for each phase volume shown, compositeswith different droplet size could be obtainedusing the different agarose microgels prepared.The micrographs for composites prepared with“no screen” microgel show a broader distributionof droplet size. This is due to the broader size dis-tribution of droplets in the agarose microgel sus-pension prepared (see Section 3.3.1).

3.3.3 - Effect of droplet size on mechanical prop-erties of maltodextrin/agarose compositesThe effect of droplet size on large deformationproperties of maltodextrin/agarose compositeswas studied for composites at different phasevolumes and modulus ratios and is shown on Fig-ures 7, 8, and 9a,b and c for, respectively, the rel-ative apparent elastic modulus, the relative truestress and strain at failure. For clarity, the behav-iour of the relative true strain at failure as a func-tion of droplet size and phase volume is shownon different graphs for different matrix concen-trations (Figure 9a, b and c).

As shown in figures 7, 8 and 9a, b and c, theeffect of droplet size on the mechanical proper-ties of maltodextrin/agarose compositesdepends on the phase modulus ratio. For com-posites with low phase modulus ratio (EAg/ESA2 =2) no effect of droplet size could be observed (inall cases, data for the three composites with dif-ferent droplet size overlap), whereas an effect ofdroplet size could be seen for composites withhigher phase modulus ratios (EAg/ESA2 = 9 and 27).This observation confirms the previous observa-tion that the filler has a larger effect if the phasemodulus ratio is high. This could explain whysome authors did not observe an effect of dropletsize on mechanical properties [5, 7].

Moreover, the effect of droplet size on themechanical properties of maltodextrin/agarosecomposites depends on the phase volume. Forcomposites with EAg/ESA2 = 9, an effect of dropletsize can only be observed (i) for the apparent elas-tic modulus at phase volume of agarose micro-gel higher than 0.3 and (ii) for the fracture prop-erties, at phase volumes above 0.4. Forcomposites with higher phase modulus ratio(EAg/ESA2 = 27), the effect of droplet size isobserved earlier (at phase volume higher than0.2). This latter observation is in accordance withthe observations of Ross-Murphy and Todd [4]and Langley & Green [6]. Ross-Murphy & Todd[4] showed that for the spherical glass/gelatin


Figure 6 (below):Confocal microscopic picturesof 30%maltodextrin / 8%agarose microgel compositeswith different sizes of aga-rose microgels (“low shear”microgel, “high shear” micro-gel and “no screen” microgel)and different phase volumes.

Figure 7 (left above):Effect of droplet size on thebehaviour of the relativeapparent elastic modulus ofmaltodextrin / agarose com-posites as a function of phasevolume and phase modulusratio.

Figure 8 (right above):Effect of droplet size on thebehaviour of the relative truestress at failure of maltodex-trin / agarose composites as afunction of phase volumeand phase modulus ratio.

composites, the size of the glass spheres has aneffect only at phase volumes higher than 0.48.

For the present study, when an effect ofdroplet size on the mechanical properties of mal-todextrin/agarose composites is observed thecomposites made with the largest droplets are thestrongest. For example, for composites with 44%agarose microgels and high phase modulus ratio,the apparent elastic modulus of the compositeswith the smallest droplets is 284 ± 4 kPa, where-as it is 483 ± 15 kPa in presence of largest droplets.The true stress at failure increases from 19.5 ± 0.7to 34 ± 0.9 kPa and the true strain at failuredecreases from 0.1042 ± 5.1·10-3 to 0.0807 ± 5·10-3.This effect contradicts other studies of the effectof glass sphere sizes on the large deformationproperties of gelatin and whey protein gels [4, 6].These authors found that smaller particlesincrease the moduli of the composites most. Therehave also been attempts to correlate the stress atfailure with the diameter of the filler particles,which show a linear relationship between either(i) the stress at failure and the reciprocal of the par-ticle diameter (for particles lower than 0.12 mm)[29] or (ii) with the reciprocal of the square root ofthe particle diameter [30]. They all found adecrease of the stress at failure with an increaseof particle size, the opposite of the present case.Our conclusions are surprising and difficult toexplain. One hypothesis could be that the differ-ing size distribution of the microgels affects theirpacking in such a way as to make the larger micro-gels have a lower maximum packing fraction, andthus higher modulus.

4 CONCLUSIONSStudies have been made of model food basedcomposites consisting of two biopolymer phas-es, maltodextrin and agarose. The materialswere produced by embedding agarose microgelparticles in a maltodextrin matrix. In contrast tothe more classic method of preparation of thecomposites used by Loret et al. [2], this processallows independent control of dispersed phasevolume, particle size and the mechanical proper-ties of either phase. The moduli and fractureproperties of maltodextrin/agarose compositeswere measured in compression as a function ofphase volume, phase modulus ratio and dropletsize. The modulus of the filler (agarose microgel)was always higher than the modulus of thematrix.

Increasing the phase volume of agarosemicrogels in maltodextrin/agarose compositeschanges the general behaviour of the compos-ites under large deformation, going from a strainsoftening behaviour similar to maltodextrin gelbehaviour (f < 28 %) to a more linear behavioursimilar to agarose gel behaviour for high con-centrations. This increase of the phase volume ofhard filler in a composite also reinforces the com-


Figure 9a,b&c:Effect of droplet size on thebehaviour of the relativetrue strain at failure of mal-todextrin / agarose compos-ites as a function phase vol-ume and phase modulusratioa) EAg/ESA2 = 27,b) EAg/ESA2 = 9 andc) EAg/ESA2 = 2.

a)

b)

c)

posite by increasing its modulus and stress at fail-ure and decreasing its strain at failure. This effectis more pronounced if the difference in modulusbetween filler and matrix is large. The effect ofphase modulus ratio on the true strain at failureis less pronounced, with a bigger decrease of thetrue strain at failure only at high phase volumefor composites with high phase modulus ratio.

Despite the multitude of models proposedto describe the behaviour of the relative modu-lus with phase volume, none was able to fit thedata correctly. However, these models describedthe effect of phase modulus ratio well. For frac-ture properties, the Ross-Murphy modelsdescribe well the behaviour of the true strain andstress at failure as a function of phase volume. Ifthe filler modulus is much higher than that of thematrix, the behaviour of the true strain at failurecan be well described by the simple relation: e =1 - ff. The Ross-Murphy model (Eq. 17) describesrelatively well the behaviour of the true stress atfailure as a function of agarose microgel phasevolume for maltodextrin/agarose compositeswith low phase modulus ratio. The introductionof the maximum packing fraction helps thedescription of the behaviour of the true stress atfailure as a function of phase volume for com-posites with high phase modulus ratio.

By including agarose microgels of differentsizes in maltodextrin gels of different strengths,the effect of droplet size on large deformationproperties of maltodextrin/agarose compositeswas studied for different phase volumes ofagarose microgels. The effect of droplet size onmechanical properties of maltodextrin/agarosecomposites depends on both the phase modulusratio and the phase volume. It was found thathigher the phase modulus ratio, the bigger wasthe reinforcement effect, i.e. higher apparentelastic modulus and the true stress at failure andsmaller the true strain at failure. This effect wasobserved at lower phase volumes for compositeswith high phase modulus ratio. For the systemwith a small modulus ratio between filler andmatrix, no effect of droplet size could beobserved. Contrary to previous studies, the com-posites with the larger droplets had the biggerreinforcement effect. They were stronger andstiffer.

These observations are very relevant for thefood industry. To develop products with new tex-tures by mixing two biopolymer gels together,

certain conditions need to be fulfilled. To see aneffect of reinforcement of the filler, the stiffnessof the filler needs to be higher than to the matrix.This modulus ratio between filler and matrix alsodetermines the phase volume at which the fillermust be included. The larger the phase modulusratio, the less filler has to be incorporated to seea reinforcement effect.

REFERENCES[1] Loret C, Frith WJ, Fryer PJ: Phase separation and

molecular weight fractionation behaviour ofmaltodextrin/agarose mixtures, Food Hydrocol-loids 19 (2005) 557-565.

[2] Loret C, Frith WJ, Fryer PJ: Mechanical and struc-tural properties of maltodextrin/agarose gelscomposites, Applied Rheology 16 (2006) 248-257.

[3] Ring S, Stainsby G: Filler reinforcement of gels,Progress in Food and Nutrition Science 6 (1982)323-329.

[4] Ross-Murphy SB, Todd S: Ultimate tensile mea-surements of filled gelatin gels, Polymer 24 (1983)481-486.

[5] Brownsey GJ, Ellis HS, Ridout MJ, Ring SG: Elas-ticity and failure in composite gels, Journal ofRheology 31 (1987) 635-649.

[6] Langley KR, Green ML: Compression strength andfracture properties of model particulate foodcomposites in relation to their microstructureand particle-matrix interaction, Journal of Tex-ture Studies 20 (1989) 191-207.

[7] Richardson RK, Robinson G, Ross-Murphy SB,Todd S: Mechanical spectroscopy of filled gelatingels, Polymer Bulletin 4 (1981) 541-546.

[8] Langley KR, Green ML, Brooker BE: Mechanicalproperties and structure of model compositesfoods, in Food Polymers, Gels, and Colloids, ed.Dickinson E, Royal Society of Chemistry, Cam-bridge (1991).

[9] Langley KR, Martin A, Ogin SL: The effect of fillervolume fraction on the fracture-toughness of amodel food composite, Composites Science andTechnology 50 (1994) 259-264.

[10] Mofidi N, Aghai-Moghadam M, Sarbolouki MN:Mass preparation and characterization of algi-nate microspheres, Process Biochemistry 35(2000) 885-888.

[11] Frith W, Norton IT: Mechanical properties ofmodel composites produced from food biopoly-mers: influence of biopolymer-biopolymer inter-facial properties, In Supramolecular and colloidalstructures in biomaterials and biosubtractes Eds.Lal M, Lillfor, PJ, Naik VM, Praskash V, Imperial Col-lege Press (1999).

[12] Frith WJ, Lips A, Norton IT: Steady shear flow prop-erties of spherical microgel suspensions, In Struc-


ture and dynamics of materials in the mesoscop-ic domain, Eds. Lal M, Mashelkar RA, Kulkarni BD,Naik VM, Imperial College Press (1997).

[13] Loret C, Frith WJ, Fryer PJ: Rheological character-isation of the gelation behaviour of maltodextrinaqueous solutions, Carbohydrate Polymers 57(2004) 153-163.

[14] Van Vliet T, Luyten T: Fracture mechanics of solidfoods, In New physico-chemical techniques forthe characterisation of complex food systems-first edition, Ed. Dickinson E, Kluwer AcademicPublishers (1995).

[15] Aymard P, Martin DR, Plucknett K, Foste, TJ, ClarkAH, Norton IT: Influence of thermal history on thestructural and mechanical properties of agarosegels, Biopolymers 59 (2001) 131-144.

[16] Lipatov Yu S: Relaxation and viscoselastic prop-erties of heterogeneous polymeric compositions,Advances in Polymer Science 22 (1977) 21-59.

[17] Chow TS: The effect of particle-shape on themechanical properties of filled polymers, Journalof Materials Science 15 (1980) 1873-1888.

[18] Ahmed S, Jones FR: A Rreview of particulate rein-forcement theories for polymer composites, Jour-nal of Materials Science 25 (1990) 4933-4942.

[19] Kerner EH: The elastic and thermo-elastic prop-erties of composite media, Proceeding PhysicSociety 69B (1956) 808-813.

[20] Nielsen LE: Dynamic mehcanical properties offilled polymers, Applied Polymer Symposia 12(1969) 249-265.

[21] Sudduth RD: Modulus evaluation of particulatecomposites using generalized viscosity model for

solutions with suspended particles, Journal ofApplied Polymer Science 54 (1994) 1243-1262.

[22] Takayanagi M, Harima H, Iwata Y: Viscoelasticbehavior of polymer blends and its comparisonwith model experiments, Memoirs Faculty Engi-neering Kyssha University 23 (1963) 1-13.

[23] Smallwood HM: Limiting law of the reinforce-ment of rubber, Journal of Applied Physics (1944)758-766.

[24] Nielsen LE: Simple theory of stress-strain proper-ties of filled polymers, Journal of Applied PolymerScience 10 (1966) 97-103.

[25] Nicolais L, Nicodemo L: Strength of particulatecomposite, Polymer Engineering and Science 13(1973) 469.

[26] Piggott MR, Leidner J: Misconceptions aboutfilled polymers, Journal of Applied Polymer Sci-ence 18 (1974) 1619-1623.

[27] Landon G, Lewis G, Boden GF: Influence of Parti-cle Size on Tensile-Strength of Particulate - FilledPolymers, Journal of Materials Science 12 (1977)1605-1613.

[28] Landel RF, Moser BG, Bauman AJ: Rheology ofconcentrated suspensions: effect of a surfactant,Proceeding 4th International Congress on Rheolo-gy (1965) 663-692.

[29] Alter H: Filler particle size and mechanical prop-erties of polymers, Journal of Applied Polymer Sci-ence 9 (1965) 1525-1531.

[30] Hojo H, Toyoshima WE, Tamura M, Kawamura N:Short- and long- term strength characteristics ofparticulate-filled cast epoxy resin, Polymer Engi-neering and Science 14 (1974) 604-609.

Applied RheologyVolume 17 · Issue 3 31412-13

volume 17 · issue 3 31412-1 - rheology · zusammenfassung: in der vorliegenden studie wird eine...

Documents