large deformation mechanical behavior of gelatin-maltodextrin composite gels

Large Deformation Mechanical Behavior ofGelatin–Maltodextrin Composite Gels

VALERY NORMAND,* KEVIN P. PLUCKNETT, STEPHEN J. POMFRET, DUDLEY FERDINANDO, IAN T. NORTON

Unilever Research Colworth, Colworth House, Sharnbrook, Bedford, MK44 1LQ, United Kingdom

Received 29 February 2000; accepted 29 August 2000Published online 18 July 2001; DOI 10.1002/app.1831

ABSTRACT: The large deformation failure behavior of gelatin–maltodextrin compositegels was assessed. All the studied compositions were selected to lie within the incom-patibility domain of the gelatin–maltodextrin phase diagram at 60°C, which producedgelatin continuous (maltodextrin included) and maltodextrin continuous (gelatin in-cluded) composites. Composite microstructural evaluation was performed using confo-cal laser scanning microscopy (CLSM). The large deformation mechanical behavior wasmeasured in tension and compression experiments. Crack–microstructure interactionswere investigated by dynamic experiments on the CLSM. The gelatin continuouscomposites exhibited pseudo-yielding behavior during tension and compression testing,and there was a significant decrease in modulus that arose from interfacial debonding.Conversely, the maltodextrin continuous composites exhibited an essentially brittlefailure behavior, and there was an approximately linear increase in stress with increas-ing strain until fracture (which occurred at significantly lower strains than for thegelatin continuous composites). The CLSM observation of the failure of the notchedsamples also demonstrated interfacial debonding in the crack path; however, thisoccurred at significantly smaller strains than for the gelatin continuous samples withminimal elastic–plastic deformation of the maltodextrin matrix. The Poisson ratio wasestimated to be close to 0.5 for these composites for all examined compositions. Com-positions corresponding to a tie line of the phase diagram were also investigated toassess the influence of the relative phase volume (for constant phase compositions) onthe failure behavior. The majority of the parameters subsequently extracted from thestress–strain curves were apparently functions of the individual phase volumes. © 2001John Wiley & Sons, Inc. J Appl Polym Sci 82: 124–135, 2001

Key words: maltodextrin; fracture; biopolymer compatibility; interface

INTRODUCTION

The small deformation mechanical behavior for agelatin (from lime hide, LH1)/maltodextrin (SA2)composite system has been extensively stud-ied.1–6 The purpose of the present work is primar-

ily to assess the large deformation behavior ofthese materials. In particular, it is important tobe able to relate failure properties and fracturepropagation mechanisms to the composite micro-structure. Preliminary observations were madeon this system by Plucknett et al.7 using confocallaser scanning microscopy (CLSM) and conven-tional dynamic tests where the gelatin-rich phasewas the matrix of the composite.

Large deformation failure was studied by anumber of authors for a range of biopolymer gels,which were structurally classified according totheir different gelation mechanisms. These in-

Correspondence to: V. Normand ([email protected]).

*Current address: V. Normand, Firmenich S.A., FFC, 7Rue de la Bergere, 1217 Meyrin 2, Switzerland.Journal of Applied Polymer Science, Vol. 82, 124–135 (2001)© 2001 John Wiley & Sons, Inc.

124

clude networks based on polymer chains (e.g.,gelatin), particle gels (e.g., casein), and gels ofdensely packed deformable particles (e.g.,starch).8 Several attempts were also made tocharacterize and model the failure behavior ofcomposite gels containing rigid9,10 and deformableparticle inclusions.11–13 These and other modelstypically assume either strongly or weakly bondedinterface conditions between the filler particlesand the matrix.

In the present study the mechanical responseof gelatin–maltodextrin composite gels was deter-mined in both tension and compression. The con-ventional dogbone-shaped samples and notchedcompact tension (CT) specimens were used fortension tests, and bulk cylindrical samples wereused in the compression studies. Both of thesetypes of test were also performed dynamically ona CSLM microscope to assess the influence ofdeformation on the gel structure. In particular,the CT technique was adopted to investigate mi-crostructure–crack interactions. The influencesof the continuous phase composition, includingthe particle size and phase volume, were investi-gated.

EXPERIMENTAL

Composite Fabrication

The gelatin was provided by SKW (Carentan,France) and was prepared by extraction from limehide (LH1; PI 5 4.7, Mn 5 83,300 Da, polydisper-sity 5 1.77, measured by GPC). Gelatin solutionswere prepared by dissolving the powder at 60°Cfor 30 min in deionized water. To prevent bacte-riological degradation of the protein, 500 ppm ofsodium azide was added. Sirius red (500 ppm)was also added to all preparations to increase thefluorescence of the gelatin (i.e., for CLSM exami-nation). When such solutions are cooled below28°C, a 3-dimensional gel network appears inwhich the gelatin chains are linked by triple he-lices.

The maltodextrin (Paselli SA2) was providedby Avebe (Foxhol, The Netherlands). It is a poly-saccharide biopolymer (a mixture of linear amy-lose and branched amylopectin) obtained afterenzymatic degradation of potato starch14 (Mn5 6.2 6 0.5 105 Da, polydispersity 5 1.45 6 0.30,measured by light scattering). The solubiliza-tion of the maltodextrin was considered to becomplete when the preparation was maintained

at 98°C for a minimum of 30 min. Orderingoccurs at temperatures below 31°C. The forma-tion of the maltodextrin gel is strongly concen-tration and time dependent.15 Maltodextrin isbelieved to form a gel of densely packed crystal-lite particles.

The composite samples were prepared by mix-ing solutions of the two biopolymers at 60°C toachieve the overall compositions indicated in thephase diagram shown in Figure 1. All of the sam-ples prepared in the current study were situatedwithin the incompatibility domain of the gelatin–maltodextrin system.5

In order to examine the influence of the phasevolume on the composite behavior (at constantphase composition), starting solutions for compos-ite construction were prepared by mixing (inequal volumes) a 24% (w/w) gelatin solution witha 24% (w/w) maltodextrin solution at 60°C underlight stirring for 5 min. This mixture was thenstored at rest in a 60°C thermostat controlledoven to achieve phase equilibration. The gelatin-rich phase creamed, and the maltodextrin-richphase sedimented. After 5 h at the equilibrationtemperature, two clear phases were observed.These phases were then decanted; by carefullyweighing proportions for each phase, different fi-nal composites could be made with differentphase volume ratios but identical phase composi-tions. In this procedure mixtures of the phasescontaining known phase volume fractions werehomogenized for 5 min under strong stirring priorto gelation.

Figure 1 The phase diagram of the system of LH1e:SA2 in 0.1M NaCl solvent at 60°C.

GELATIN–MALTODEXTRIN COMPOSITE GELS 125

Tension Test Sample Preparation

Solutions prepared as above were subsequentlypoured between parallel glass plates (separatedby 1.4-mm spacers and covered with hydrophobicpaper to prevent the gelled solutions from stick-ing to the glass) preheated to 60°C in a water bathand protected from water contact by a plastic bag.The plates were then quenched rapidly to 9°C byimmersion in a second water bath and stored for24 h at 5°C.

Compression Test Sample Preparation

The solutions were poured into cylindrical molds(12.2-mm length and 12.5-mm diameter), the sur-face of which was previously covered by vacuumgrease to prevent gel sticking. Samples were thenrapidly cooled to 10°C to avoid large-scale phaseseparation.

Mechanical Testing and MicrostructureCharacterization

Compression and tensile tests were conducted us-ing an Instron Universal testing machine (model4501). Samples were equilibrated at 10°C prior toand during mechanical testing using a tempera-ture controlled cabinet.

Tension Tests

A dogbone-shaped specimen geometry was chosenfor the tensile tests, and samples were cut fromthe 1.4 mm thick gel sheets using a suitablyshaped cookie cutter. The test specimens werethen gripped on the Instron using double-sidedtape. Only samples that failed within the gaugelength were used for calculation of mechanicalproperties. Testing was conducted using a dis-placement rate of 50 mm/min. True failure stress(st) and true failure strain (et) was subsequentlycalculated using the following equations:

st 5F~L0 1 DL!

A0L0(1)

«t 5 lnSL0 1 DLL0

D (2)

where F is the load, L0 is the initial gauge length,DL is the displacement, and A0 is the cross-sec-tional area.

Compression Tests

The compression tests made under lubricatedconditions (dodecan) were also conducted at 50mm/min. The true stress (sc) and true strain (ec)were calculated using the following equations:

sc 5FH

A0H0(3)

«c 5 lnSHH0

D (4)

where H0 is the initial sample height and H is thesample height after deformation.

CLSM Characterization

The microstructures of the fabricated compositeswere assessed using a CLSM microscope. For themicrographs presented in this study the gelatin-rich phase was light in color while the maltodex-trin-rich phase was dark, which was due to pref-erential staining of the gelatin by Sirius Red,which exhibits fluorescence under laser illumina-tion. Large deformation mechanical tests, includ-ing observation of failure, were also performed onthe CLSM microscope using a Minimat tension/compression testing stage. Two tension test geom-etries were utilized for these experiments: theconventional tensile test described above and anotched CT test. For the CT experiments thenotched samples were subjected to small steppedtensile displacements (typically 0.1-mm steps at adisplacement rate of 2 mm/min) while observingthe notch tip on the CLSM. All CLSM visualiza-tion experiments were performed at 22 6 1°C. Itwas generally observed that the notch eventuallybegan to grow stably across the test specimen,after which further tensile displacements weregenerally unnecessary. Crack advance could thenbe followed directly on the CLSM microscope andrecorded dynamically on video. In addition to thetension experiments, cylindrical samples werecompressed between two lubricated glass platesand observed from above using the microscope tovisually assess the deformation behavior in com-pression.

RESULTS AND DISCUSSION

Gelatin Continuous Composite Behavior

A typical set of Instron load–displacementcurves, obtained for the gelatin continuous com-

126 NORMAND ET AL.

position III (see Fig. 1) is shown in Figure 2.These observations were generally characteristicof gelatin continuous samples, and the large de-formation failure curves showed three distinctregions.7 The initial small displacement response(,5 mm displacement, equivalent to 8% truestrain) corresponded to the linear elastic regionwith constant elastic modulus. This was followedby a region of decreasing elastic modulus withpseudo-ductile behavior and finally the samplefractures at large extensions (i.e., ,50% strain).

Three different composite compositions wereextensively investigated: composites I, II, and IIIwere 8% SA2 and 9.3% LH1e, 12% SA2 and 9.3%LH1e, and 12% SA2 and 12% LH1e in 0.1M NaCl,respectively. The significant difference betweenthese composites was the ratio of the stiffness ofthe included phase relative to that of the matrix.In composite system I the matrix modulus washigher than the particle modulus. In system II theparticle and matrix moduli were similar, while insystem III the modulus of the particles was largerthan that of the matrix. A comparison of the com-posites in which gelatin was the continuous phasewith pure gelatin behavior was already discussedin a preliminary article.7

It was apparent for these systems that thestandard deviation error in the elastic moduluswas small when compared to the error observedfor the other failure properties. The elastic mod-ulus can be viewed as a nominally intrinsic ma-terial property (dependent upon composition andprocessing and test conditions). However, the fail-ure stress was largely dependent upon the largestflaw size and the flaw size distribution in any

particular test specimen. For the immisciblephase-separated composites examined in thepresent study, the strength-controlling flawswere likely either large second phase particle in-clusions, entrapped air bubbles, or surface andsubsurface damaged regions caused by the speci-men cutting procedure. This emphasized thestrong influence of the microstructure on thelarge deformation behavior of such materialswhen compared to that in the linear viscoelasticsmall deformation region.

A CLSM image of the microstructure of compo-sition I after failure is presented in Figure 3(a)and the edge of the fracture surface is shown onthe right-hand side. The maltodextrin particlesize was relatively small (10–20 mm), and the sizedistribution was not very broad. Evidence of re-sidual strain was seen around the droplets of theincluded phase, mostly oriented in the directionnormal to the fracture. The fracture surface was

Figure 3 A CLSM microscope image of the edge of afracture surface for compositions (a) I and (b) II (gelatinwhite, SA2 dark).

Figure 2 The tensile load–displacement curves ob-tained for composition III with rapid cooling (smallinclusions, curves a and b) and slow cooling (large in-clusions, curves c–g,).


irregular, debonding of SA2-rich particles wasseen, and the particle shape remained intact onthe fracture surface.

The microstructure of composition II after fail-ure is shown in Figure 3(b). The particle size wasgenerally larger than for composition I, as mightbe expected because of the higher SA2 phase vol-ume in this sample, which promoted droplet co-alescence. However, small particles were stillseen (10 mm). The load–deflection behavior of thissample was shown in Figure 2. In Figure 3(b)interfacial debonding between the particle andthe matrix is apparent, creating voids around theparticle, predominantly in the direction of theapplied strain. Qualitatively, only the largest par-ticles exhibited this debonding behavior once thesample was relaxed after failure.

When the sample was cooled down at a higherrate, the maltodextrin particle size was signifi-cantly reduced for composition II (Fig. 2, curves a

and b). It was notable that the initial elastic mod-ulus, calculated in the linear elastic region of theload deflection curve, remained consistent for allsamples, regardless of the included particle size(Fig. 2). A similar observation was reported byBrownsey et al.13 for a gelatin-based system rein-forced with Sephadex beads. However, althoughthe modulus was essentially unchanged with areduction in particle size, the failure load (stress)and displacement (strain) both increased dramat-ically (Fig. 2). Particle–matrix debonding was lessapparent in the sample with smaller particle size(after failure and relaxation) as supported by ear-lier work.16 In addition, the shape of the particleswas qualitatively less spherical and they ap-peared to be stretched in the direction of the ap-plied strain, which indicated a ratio of moduliclose to 1 between the particle and the matrix.

Composition III exhibited tensile behavior sim-ilar to system II. Compression tests were alsoconducted under the CSLM microscope betweentwo glass plates. These revealed that debondingoccurred in the equatorial plane of the particle[Fig. 4(a,b)]. In this system debonding appearedat a larger strain than under applied tensilestress. For the large deformation tests (Fig. 5) theshape of the curve under compression was similarto that obtained under tension, exhibiting a linearresponse at low strain values (below 5%), then aductile region, followed by fracture of the sampleat large strains. However, the yield strain (i.e.,strain at the onset of debonding) under compres-sion was roughly twice the yield strain undertension. This behavior can be expected for an

Figure 4 A CLSM image of the compression test un-der the microscope with (a) an uncompressed sampleand (b) compressed at around 30% true strain.

Figure 5 A comparison between tension and com-pression for system III (12% LH1e:12% SA2 in 0.1MNaCl solution). The true strain and stress were calcu-lated according to eqs. (1)–(4).

128 NORMAND ET AL.

incompressible material (i.e., Poisson ratio 5 0.5)and the yield stress was roughly the same, whichindicated that even under compression the failurewas tension controlled. It was also noted that thefracture strain under compression was 2–3 timesgreater than that observed for tension. TheYoung’s moduli measured in the early stages ofthe deformation process (i.e., strains , 15%) wereessentially identical for the tension and compres-sion tests.

Maltodextrin Continuous Sample

The maltodextrin continuous samples exhibited aclassical brittle failure response in tension7 witha linear increase in displacement with increasingload, followed by sample failure. Clearly, theshape of the tensile curve was very different thanthat in the previous case (i.e., gelatin continuous)where pseudo-ductile behavior was noted.

Although the failure stresses for the maltodex-trin continuous samples were comparable to thoseof the gelatin continuous materials, the failurestrains were an order of magnitude lower, high-lighting their brittle behavior, and linear elasticbehavior occurred up to fracture.

The CLSM images of the microstructures ofcompositions IV and V after failure are shown inFigure 6(a,b). The gelatin inclusions were ;100mm in diameter for compositions IV and V, and noobvious sign of particle percolation was noted.

In Figure 6(a) (composition IV) the fracturesurface exhibits some distinct differences fromthe gelatin continuous samples. There are voidswhere a particle was detached, and the particlethat remained with the sample appears to be un-perturbed by the fracture path. The crack pathgoes around the particles (i.e., there is debond-ing), which again suggested that the interfacebetween the maltodextrin matrix and the gelatindroplets was weak. However, although this be-havior was somewhat similar to the gelatin con-tinuous materials, it is clear that there was littleor no particle–matrix debonding away from theactual crack, which was anticipated from the sig-nificantly lower failure strains observed for themaltodextrin continuous samples.

Figure 6(b) demonstrates a similar cross sec-tion of the fracture surface of composition V. It isapparent that once a gelatin droplet was partiallypulled out from the matrix, it showed a residual(permanent) strain. This behavior was noted con-sistently for gelatin particles directly in the crackpath; however, such apparent plastic deformation

was not observed away from the crack. Tensilefracture studies of pure monophase gelatin gelsdemonstrated that apparent unrecoverable plas-tic deformation can occur at high strains, whichwas attributed to structural alignment duringstraining.17

To further explain the fracture behavior of themaltodextrin continuous composites, notched CTtests were performed dynamically on the CLSMmicroscope using a Minimat stage. The resultsare presented in Figure 7 for composition IV; thesame protocol was followed as adopted for a gel-atin continuous sample discussed in Plucknett etal.7 The displacement necessary to initiate crackgrowth from the original notch was 0.4 mm,which was significantly less than required for thegelatin continuous samples (several millimeters).It was readily apparent that the shape of thecrack was considerably different than the gelatin

Figure 6 A CLSM image of the edge of a fracturesurface for compositions (a) IV and (b) V (gelatin lightinclusions, maltodextrin dark matrix). Residual straincan be observed in particles at the fracture surface.


continuous samples, which exhibited a veryblunted crack tip morphology (reminiscent of duc-tile failure in metals); the tip morphology for themaltodextrin continuous composites was sharp(cf. ductile metals and brittle ceramics18). In Fig-

ure 7(a) the crack tip is close to a gelatin particle,which apparently remains unstressed comparedto gelatin continuous samples.7 In reality thecrack tip was a region of stress concentrationwhen the composite was under applied load (i.e.,

Figure 7 CLSM images of a notched CT test for composition V (slow cooling, largeparticles). Debonding of an included gelatin droplet in the continuous maltodextrinmatrix is apparent. The bright spot corresponds to the crack tip, and the debondingevent is highlighted with the white box.

130 NORMAND ET AL.

applied displacement). Further crack propagation[Fig. 7(b)] resulted in initial interfacial debondingbetween the spherical particle and the continuousmaltodextrin matrix. This was revealed by a lossof contrast in the particle because air was presentat the interface (shown in lower left side). InFigure 7(c,d) the crack propagates around the leftside of the particle. As the crack advances further[Fig. 7(e,f)] the particle is totally detached fromthe matrix and pulled out of the left side of thecrack, and a matching hole is retained on thatside with the particle remaining “bonded” to theright side of the crack.

When the concentration of gelatin particleswas higher, healing (or bonding) of neighboringparticles occurred during gelation of the malto-dextrin-rich matrix phase that occurred for somecompositions after gelation of the particles. Thisis apparent in Figure 8(a,b), where crack propa-

gation is retarded by ligaments of gelatin dropletsbound to each other in the crack wake. This be-havior may be viewed as a toughening mecha-nism for the brittle maltodextrin matrix (in amanner analogous to ductile metal toughened ce-ramics). Figure 8(a–c) shows a gelatin ligamentthat is stretched across the crack wake until it isfinally pulled out from one side of the crack [Fig.8(d)]. The strain in this ligament prior to pulloutwas significantly greater than 100%. In Figure8(b) a new bridging ligament appears (white ar-row), which is then followed by another one [Fig.8(c)].

When a large notch was initially cut into themaltodextrin continuous samples, rapid samplefracture occurred. However, fine gelatin liga-ments still bridged the crack with the samplebeing in two pieces. With such rapid failure of thesample in this instance it was impossible to de-

Figure 8 CLSM images of composition IV during the notched CT test (edge of theplate, small particles). In the insets there is (a) evidence of coalescence of gelatinparticles, (b,c) gelatin ligament formation and stretching, and (d) relaxation afterligament pull-out. (b–d) Further ligament bridges are forming.


termine an accurate strain rate for the ligamentdeformation. However, with apparent strains ofseveral hundred percent and failure occurring inless than 1 s, a strain rate in excess of 102 to 103

s21 can be estimated. It is interesting to note thatthe strains observed in these ligaments were sig-nificantly higher than those noted during largedeformation testing (under shear and compres-sion) of pure gelatin, typically 150–200%.19

For deformation under compression, maltodex-trin continuous samples did not exhibit ductilebehavior, indicating that the matrix deformedand fractured before debonding between the gel-atin particle and maltodextrin matrix could occur.However, as shown in Figure 9 for system VI(33% included gelatin phase volume), for a similarYoung’s modulus, fracture of the sample occurredunder compression at a strain twice the valueobtained under tension and for a stress valuebetween 1.5 and 2 times the value obtained fortensile tests.

Effect of Phase Volume Fraction Ratios on Tensionand Compression Failure Behavior

The effect of the phase volume fraction at con-stant phase composition was also investigated un-der compression and tension. Results were ob-tained for one tie line of the phase diagram, whichcorresponded to the 12% LH1e:12% SA2 in 0.1MNaCl system (III in Fig. 1). All composites wereprepared by remixing the two separated phases toobtain the desired phase volumes, as described in

the Experimental section. Therefore, the compo-sitions and the characteristics of the two phasesand the interfaces between the two phases wereidentical from one sample to another (assumingno water repartition on gelling), and only thephase volume and the particle size varied. Theresults found for the elastic moduli were consis-tent with the simple isostrain model,20 whateverthe nature of the continuous phase (Fig. 10).

The stress at failure as a function of the strainat failure for samples tested in tension and com-pression is shown in Figure 11. It can be seen thatthe strain at failure in compression was twicethat in tension for all the samples. Compressiontests may also be analyzed in terms of biaxialextension, the applied compression resulting intension along the normal axes. For a materialwith a Poisson’s ratio of 0.5, the tension strain

Figure 9 A comparison between tension and com-pression for a maltodextrin continuous sample with33% included particles of the LH1e-rich phase. Thetrue strain and stress were calculated according to eqs.(1)–(4).

Figure 10 The Young’s modulus evolution with thephase volume of the maltodextrin-rich phase and acomparison with the Isostrain model.

Figure 11 The envelope of failure for the composites:(F, ■) compression and (E, h) tension.

132 NORMAND ET AL.

along the normal axes are half that applied incompression. The data shown in Figure 11 there-fore suggests that the failure of the samples oc-curred in tension and that the Poisson’s ratio ofthese materials was close to 0.5.

The stress at failure of the samples followed adifferent trend. For the maltodextrin continuoussamples the stress at failure in compression wastwice that of failure in tension. However, for gel-atin continuous samples the stresses at failure incompression and tension were very similar. Thismay be explained by considering the failure of thesamples to be strain dependent rather than stressdependent. For the maltodextrin continuous sam-ples the stress increased linearly with the strainin tension right up until failure, so a doubling ofstrain led to a doubling of stress, which is evidentin Figure 9 and reflected in the trend shown inFigure 11. However, for a gelatin continuous sam-ple the stress varied little with strain above theyield point (Fig. 2); so if the strain was doubledthe stress remained almost the same. Hence, inboth systems the failure strain in compressionwas twice that in tension, and the difference infailure stress was merely a consequence of thedifferent stress–strain functions.

However, the ef and sf fracture characteristicswere strongly dependent on the continuous phasebehavior and could not be treated directly as afunction of the phase volume, because the natureof the matrix had an important influence (ductilewhen LH1e was continuous and brittle when SA2was continuous). Therefore, two separated do-mains must be considered according to the com-position of the continuous phase.

For a rubberlike matrix where filler spheresare embedded with no adhesion between the ma-trix and filler, Nielsen21 suggested two funda-mental relationships between the relative stressand the relative strain at failure with the phasevolume of the filler:

sf~filled!

sf~unfilled!5 1 2 ffiller

2/3 (5)

«f~filled!

«f~unfilled!5 1 2 ffiller (6)

In Figure 12(a,b) these formulas are applied forthe gelatin-rich continuous composites. Unfortu-nately, according to this model, the size of theparticle was not taken into account. The tenden-cies seen in these figures were however in good

general agreement with eqs. (5) and (6), becausethe stresses and strains at failure decreased whenthe filler phase volume increased.

For the maltodextrin continuous compositesthe tendencies were the opposite. When the phasevolume of the gelatin increased, leading to greatercocontinuity of the structure, the strain and thestress at failure both increased. This phenomenonis largely believed to be due to the characteristicsof the maltodextrin-rich matrix, which fails beforeany significant debonding can occur. It was there-fore interesting to consider the effect of the pres-ence of gelatin droplets in a maltodextrin matrixand, more particularly, on the work necessary to

Figure 12 (a) The relative stress at failure as a func-tion of the phase volume of filler for gelatin continuouscomposites: (■) compression, (F) tension, and (- - -)1 2 fm

2/3. (b) The relative strain at failure as a functionof the phase volume of filler for gelatin continuous com-posites: (h) compression, (E) tension, and (- - -) 1 2 fm.


break the matrix. The energy of fracture can becalculated by definition as follows22:

Wf 5 E0

f

si d« (7)

Even though the gelatin–maltodextrin interfaceis weak7 and the fracture follows a route aroundthe gelatin particles, the energy needed to frac-ture samples including gelatin particles increaseswith the phase volume of the gelatin (Fig. 13). Apossible reason for the increase in the energyrequired for fracture is the energy required fordeforming the bridging gelatin particles behindthe crack tip (i.e., in the crack wake, see Figs, 7, 8).

CONCLUSION

The matrix–inclusion interface in gelatin–malto-dextrin composites plays an extremely importantrole in their failure behavior. All samples pre-pared exhibited spherical inclusions of one phaseembedded in a continuous matrix of the other.Gelatin continuous systems exhibited lowerstrength but were more resistant to fracture thanthe maltodextrin continuous systems (e.g., theelastic modulus was lower for gelatin continuoussamples, but the failure strains were an order ofmagnitude greater). The shapes of the stress–strain (or load–displacement) curves were differ-ent, depending upon the continuous phase na-ture. The maltodextrin continuous samples were

essentially brittle in nature, while the gelatincontinuous samples exhibited a pseudo-plasticbehavior (with an apparent yield stress). Thislatter phenomenon was attributed to interfacialdebonding of maltodextrin particles from the gel-atin matrix, which was observed when conductingdynamic tests on the CSLM microscope.

A comparison between compression and ten-sion behavior was made and it highlighted thefact that the strain at failure in tension was twiceas much as the strain at failure in compression.As a consequence, a Poisson ratio close to 0.5 wasfound, whatever the nature of the continuousphase in that system. The observed phase volumeeffect on the failure properties at a constant phasecomposition could correspond to the Nielsen modelfor nonadhesive filler particles in a rubberlikematrix, but improvement in this model is neededif the included particle size effect on the failureproperties is to be described. Low adhesion be-tween particles and matrix must be a criterion fora debonding phenomenon.

The authors would like to thank Dr. W. J. Frith andProf. A. H. Clark for valuable discussions and com-ments.

REFERENCES

1. Kasapis, S.; Morris, E. R.; Norton, I. T.; Clark, A. H.Carbohydr Polym 1993, 21, 243.

2. Kasapis, S.; Morris, E. R.; Norton, I. T.; Gidley,M. J. Carbohydr Polym 1993, 21, 249.

3. Kasapis, S.; Morris, E. R.; Norton, I. T.; Brown,C. R. T. Carbohydr Polym 1993, 21, 261.

4. Kasapis, S.; Morris, E. R.; Norton, I. T.; Clark, A. H.Carbohydr Polym 1993, 21, 269.

5. Normand, V.; Pudney, P. D. A.; Aymard, P.; Nor-ton, I. T. J Appl Polym Sci 2000, 77, 1465.

6. Loren, N.; Langton, M.; Hermansson, A. M. FoodHydrocolloid 1999, 13, 185.

7. Plucknett, K. P.; Normand, V.; Pomfret, S. J.; Fer-dinando, D. P. Polymer 2000, 41, 2319.

8. van Vliet, T.; Walstra, P. Faraday Discuss 1995,101, 359.

9. Ross-Murphy, S. B.; Todd, S. Polymer 1983, 24,481.

10. Gao, Y. C.; Lelievre, J. Polym Eng Sci 1994, 34,1369.

11. McEvoy, H.; Ross-Murphy, S. B.; Clark, A. H. Poly-mer 1985, 26, 1483.

12. McEvoy, H.; Ross-Murphy, S. B.; Clark, A. H. Poly-mer 1985, 26, 1493.

13. Brownsey, G. J.; Ellis, H. S.; Ridout, M. J.; Ring,S. G. J Rheol 1987, 31, 635.

Figure 13 The work of fracture of the continuousphase as a function of the phase volume of the contin-uous phase for maltodextrin continuous composites.

134 NORMAND ET AL.

14. Defloor, I.; Vandenreyken, V.; Grobet, P. J.;Delcour, J. A. J Chromatogr A 1998, 803,103.

15. Durrani, C. M.; Prystupa, D. A.; Donald, A. M.;Clark, A. H. Macromolecules 1993, 26, 981.

16. Gent, A. N. J Mater Sci 1980, 15, 2884.17. Bigi, A.; Bracci, B.; Cojazzi, G.; Panzavolta, S.;

Roveri, N. Biomaterials 1998, 19, 2335.18. Atkins, A. G.; Mai, Y. W. Int J Fracture 1986, 30,

R49.19. Bot, A.; van Amerongen, I. A.; Groot, R. D.; Hoek-

stra, N. L.; Agterof, W. G. M. J Chim Phys PhysChim Biol 1996, 93, 837.

20. Takayanagi, M.; Harima, H.; Iwata, Y. Membr FacEng Kyushu Univ 1963, XXIII, 3.

21. Nielsen, L. E. J Polym Sci Polym Phys Ed 1979, 17,1897.

22. van Vliet, T.; Luyten, H.; Walstra, P. In Food Col-loids and Polymers, Stability and Mechanical Prop-erties; Dickinson, E., Walstra, P., Eds.; SpecialPublication 113; Royal Society of Chemistry: Cam-bridge, U.K., 1993; p 175.


large deformation mechanical behavior of gelatin-maltodextrin composite gels

Documents