Un

co

rre

cte

d P

ro

of

RHEOLOGICAL CHARACTERIZATION OF CONCENTRATED NANOCLAY DISPER-

SIONS IN AN ORGANIC SOLVENT

FRANCESCA LIONETTO, ALFONSO MAFFEZZOLI

Department of Innovation Engineering, University of Salento, via Monteroni, 73100 Lecce, Italy

* Email: francesca.lionetto@unile.itFax: x39.0832.297240

Received: 15.5.2008, Final version: 3.12.2008

Abstract:

Nanoclay dispersions in organic solvents are widely used in cosmetics for a variety of gels and creams, whose propertiesdepend on the powder content and the processing method. The control of the shear applied during processing is thereforeessential for achieving the required properties. This study demonstrates the utility of applying rheological measurements forcharacterizing cosmetic products based on nanoclays and relating their viscoelastic properties to end-use performances. Inparticular, a rheological characterization of bentonite dispersions in isododecane at different clay content and shear historyis presented. For each inorganic content, both mixed samples and samples subjected to several calendering runs were stud-ied. The effect of shear and clay content on the viscoelastic properties was investigated by a combination of oscillatory shearexperiments under small-deformation conditions and by X-Ray diffraction. The tested samples showed a gel-like behaviourwith a final structure depending on the applied shear stress. By increasing the inorganic content in the dispersion, a reduc-tion in the gel stability to a further shear application was observed. Two models, developed for colloidal gels, were used to fitthe rheological results enabling to evaluate the microstructure and the degree of dispersion of the tested samples and torelate the colloidal structure to the elastic properties.

Zusammenfassung:

Résumé:

Key words: nanoclay, gel, rheology, viscoelastic properties, cosmetics, calendering

1 INTRODUCTIONOrganically modified layered silicates or nanoclays are extensively used in pharmaceutical and cosmetic formu-lations [1 - 6] and in a wide range of applications, such as in polymer nanocomposites, in paints, inks and greas-es as rheological modifiers, in adsorption of toxic gases, in effluent treatment and drug delivery carrier [7-12], etc.Bentonite is one of the most used layered silicates belonging to smectite group clays, which are also known as2:1 phyllosilicates, the most common of which are montmorillonite and hectorite. The crystal lattice of smectitegroup clays consists of two-dimensional 1 nm thick layers, which are made up of two tetrahedral sheets of silicafused to an edge-shaped octahedral sheet of alumina. The lateral dimensions of these layers vary from 30 nm toseveral microns depending on the particular silicate. Stacking of the layers leads to a regular Van der Waals gapbetween them, called interlayer or gallery. Because of the relatively weak forces existing between the layers (dueto the layered structure), intercalation of various molecules, and even polymers, is easy [7 - 9]. Bentonite has ahigh swelling ability in water and exhibit different rheological behaviours, liquid and viscoelastic gel, dependingon the clay content, ionic strength, pH, etc. [9 - 12].

In pharmaceutical and cosmetic industry, nanoclays are finding extensive use in the production of gelsand creams, obtained as colloidal dispersions of clay powders in a given solvent. The processing methods usual-ly apply shear to delaminate the large stack of silicate nanoplatelets into single layers or groups of layers. One ofthe most used techniques is the calendering process, which forces the material to pass through the gap betweenthree or four rotating cylinders. Unlike what happen during shaping of thermoplastic melts, the calendering ofcosmetic formulations occurs at room temperature without controlling the temperature surface of the rolls.Moreover, variations in gap size due to roll dimensions and the roll distortion due to high pressure developing inthe gap, will change the pressure profile in the material. Small differences in the locally applied shear or temper-ature can drastically affect the viscoelastic properties and the rheological behaviour of the material. This can dra-

To appear as:Appl.Rheol.19 (2009) 23423

Un

co

rre

cte

d P

ro

of

matically damage the end-use performance properties of the cosmetic products.Rheology, both in stationary and dynamic mode, is the most appropriate technique for characterizing

the cosmetic formulations. As demonstrated by Yao et al. [15] and Hamer et al. [16], simple and fast measure-ments can be implemented for predicting the storage stability, the effects of the formulation on consistency andsensory assessment and the rheological behaviour under manufacturing conditions. The knowledge of the keyfactors affecting the stability and rheological behaviour of clay-solvent dispersions is of fundamental impor-tance for understanding the mechanism of the property enhancement by shear application. While the rheolog-ical behaviour of bentonite-water dispersions has been widely studied [8 - 9, 17-21], a systematic study of the vis-coelastic properties of organic-inorganic hybrid materials obtained by dispersing bentonite in organic solventsis still lacking.

This paper presents the results of the rheological characterization of concentrated bentonite dispersionsin an organic solvent, which are widely used for the preparation of cosmetic nanofilled gels. The effect of the claycontent and the shear applied by the dispersion method (calendering) on the dynamic viscoelastic properties hasbeen investigated. Results obtained by small-amplitude oscillatory shear tests have been interpreted on thebasis of some models, developed for colloidal gels, to give information about the degree of dispersion and thestructure of the systems under investigation.

2 EXPERIMENTALIn this work an organically modified hectorite, supplied by Elementis Specialties with the commercial name ofBentone 38V was used as received. The clay powder had a density value of 1.7 g/cm3 and a cation exchange capac-ity (CEC) value of 98 meq/100g.

The dispersions with different clay content (11 and 17% by weight, corresponding to 5.1 and 8.3% by vol-ume) were prepared according to a protocol reported in the technical datasheet by the supplier. The bentonitepowder was first dispersed thoroughly in an organic solvent (isododecane) by stirring at 60°C. After 10 minutesa 95/5 solution of ethanol/water was added under agitation to promote and stabilise the dispersion, which wasfurther stirred at 60°C for 5 hours. After reporting the batch to room temperature, some samples were loaded ina three-roll calendering mill (EXAKT 50, Exakt Technologies Inc.) and subjected to one or two calendering runs.Isododecane and ethanol were supplied by Brenntag S.p.A. and Aldrich, respectively, and used as received. Thedenomination of the studied samples as a function of the shear history and clay content is reported in Tab. 1.

Dynamic mechanical analysis was performed at 25°C using a strain controlled rheometer (ARES, TAInstruments) equipped with a parallel plate geometry (25 mm diameter, 1 mm gap). The absence of slip effectswas confirmed by comparing the results obtained using parallel plates with smooth geometry and modifiedgeometry with sandpaper (220 Ìm roughness grit). No significant differences were observed, as found also byMobuchon et al. [22] for dispersions in a non-polar solvent. The rheometer was used in the oscillatory mode per-forming strain sweep tests (at a fixed frequency of 1 Hz) and frequency sweep tests (at a constant strain of 0.05%within the linear viscoelastic region). Since bentonite dispersions are quite sensitive to the deformation history,particular attention was devoted to annul the stress induced during sample loading. Samples were carefullyloaded on the lower rheometer plate and the upper plate was slowly let down, then a preshearing at 20 s-1 for10 minutes was applied. Before starting the measurement, the samples were kept at rest for 10 minutes in orderto allow the material to recover its initial structure.Wide angle X-Ray scattering (WAXS) measurements were car-ried out on a Philips PW 1729 Diffractometer using a Cu Ka radiation with wavelength l = 0.154 nm.

3 RESULTS 3.1 STRUCTUREX-ray Diffraction analysis has been carried out to quantify the widening of the gallery height (d-spacing)between adjacent silicate platelets due to the intercalation of the isododecane molecules between the galleries.The effect of the processing route (mixing and calendering) on the X-Ray diffraction patterns of the bentonitedispersions is shown in Figs 1 and 2 for samples with 11% and 17% by weight of nanoclay content, respectively. Forcomparison purposes, Figs 1 and 2 also report the spectrum of organically modified bentonite powder, tested asreceived from the supplier. The powder spectrum shows a pronounced peak at 3.4° and a small one at 7°. The firstpeak, corresponding to a spacing distance of 2.58 nm, indicates the intercalation of the organic modifier mole-cules between the bentonite lamellas [23], while the second peak can be related to a second order diffraction.

The spectra of the mixed and calendered samples with low clay content (curves “A-NC” and “A-C1” inFigure 1) are characterized by the absence of any diffraction peak and the presence of a broad amorphous band.

To appear as:Appl.Rheol.19 (2009) 23423

Un

co

rre

cte

d P

ro

of

The spectrum of the twice calendered sample (A-C2) is not reported in Fig.1 because it is perfectly overlapped tothat of once calendered sample (A-C1). The obtained spectra are a first indication of the occurred delaminationof the large stacks of silicate nanoplatelets into single layers or, more likely, tactoids of a small number of layers.Nevertheless, the absence of any difference among the diffraction patterns of the samples subjected to differ-ent shear history (A-NC, A-C1, A-C2) and, likely, with different intercalation degree, is due to an intrinsic limit ofthe measuring technique. Wide angle X-Ray scattering is a common method to prove the formation of interca-lated nanocomposites but it cannot provide useful information when the sample gallery height is above the res-olution limit of the diffractometer (about 6 nm) [24]. Increasing the clay content, a small shoulder at 3.4° is stillobserved only in the spectrum of the mixed sample (“B-NC” curve of Fig.2). The more likely explanation is thatfor more concentrated dispersions the complete intercalation of organic solvent molecules between the nan-oclay galleries can be achieved only by calendering, as demonstrated by the absence of any diffraction signalbelow 2q = 7°. No differences are observed between the spectra of the once (B-C1) and twice (B-C2) calenderedsamples. Any eventual further increase of d-spacing arising by subsequent calendering cannot be distinguishedbecause in both cases the platelet basal spacing is beyond the instrument resolution limit.

3.2 LINEAR VISCOELASTICITYIn order to draw structural and rheological information from the nanofilled gels, dynamic mechanical analysishas been performed. Bentonite dispersions are known to be thixotropic, that is the values of the moduli and themaximum strain depend on the preshear history and time [25]. To achieve reliable results on the linear viscoelas-tic response, the samples have been subjected to the same mechanical history and rest time after loading inorder to minimize the loading effects. From preliminary time sweep experiments at low strain amplitude(0.05%), the viscoelastic moduli have been observed to be constant with time after about 600s. A waiting timeof 600s has been used for all the strain sweep and frequency sweep experiments below reported.

The strain sweep measurements for the dispersions with different clay content and shear history arecompared in Fig. 3. A fixed frequency of 1 Hz has been kept constant during the experiment, while the strain hasbeen increased. The storage modulus (G’) is at least 10 times higher than the loss modulus (G’’) that, for sake ofclarity, is not reported. This indicates that the bentonite dispersions show a gel or solid-like behaviour in the lin-ear viscoelastic region probably resulting from the formation of a three-dimensional network involving weakbonding forces between the organic and inorganic phase.

During the strain sweep measurements, the storage modulus G’ remains constant as far as the samplestructure is maintained. When the gel intermolecular forces are overcome by the oscillation stress, the samplebreaks down and the modulus falls. The maximum strain (gcrit), to which the linear viscoelastic region extends,has been determined as the value above which G’ decreases more than 10% of the maximum value. As can beclearly observed in Fig. 3 and from the values reported in Tab. 2, under the same shear history, an increased claycontent leads to a very high increment in the storage modulus (one order of magnitude) and to a decrease of thelinear viscoelastic region (gcrit). The higher the clay content, the lower the maximum shear strain that can beapplied to the system keeping the condition of linear viscoelasticity.

Under the same clay content (Figure 4), with increasing the number of calendering runs the mechanicalproperties of each dispersion increase while gcrit decreases. However, an increased shear stress may produce gelinstability in the sample with the highest nanoclay content. As shown by the curve “B-C2” in Figure 3, a reduc-tion of about 40% in the G’ value is observed for the sample B after the second calendering run.

The frequency sweep curves, reported in Figure 5, have been obtained using a very low strain value (0.05%) within the linear viscoelastic region. For each concentration, G’ is larger than G’’ in the whole experimentalfrequency range showing that the elastic behaviour of bentonite dispersions is dominant compared to the vis-cous one. The viscoelastic moduli show a slight frequency dependence, which increases with clay content andcalendering number. Both storage and loss moduli reveal similar trends (only storage modulus is reported). Thelinearity of the shear modulus with frequency suggests that the studied dispersions may be considered in theused frequency range as gel-like networks.

In order to better investigate how the dispersion method affects the rheology, three consecutive fre-quency sweep tests have been performed on the same sample in a parallel plate configuration (Figures 6 and 7).Before the second and the third frequency sweep test, the sample has been subjected to a pre-shear conditionconsisting in a stationary rotation at a shear rate of 20 s-1 for 10 minutes. The shear rate value has been chosen,after some preliminary tests, as the maximum rate at which gel fracture does not occur between the rheometerplates. Although the shear rate applied before the measurement is lower than that applied during processing,the aim of the experiments was to affect the orientation of the aggregates reproducing at a laboratory scalewhat happens during industrial calendering.

To appear as:Appl.Rheol.19 (2009) 23423

Un

co

rre

cte

d P

ro

of

The structure becomes increasingly elastic with the repeated frequency sweep runs, as observed fromthe enhanced G’ values. For samples with lower bentonite content (Figure 6) this effect is more evident for thejust mixed gel, which is less structured than the twice calendered gel. In the mixed sample with higher clay con-tent (Figure 7), the increase of G’ with shear is lower than that observed in Figure 6. Moreover, the twice calen-dered dispersion evidences a decrease of G’ with increasing shear applied by the consecutive frequency sweepruns. This result definitely confirms that at higher clay content the gel-like structure may be more easilydestroyed by any further shear application after simple mixing.

4 DISCUSSIONThe dynamic mechanical results have evidenced that both composition and dispersion methods greatly affectthe viscoelastic properties of the studied gels, which present a solid-like behaviour. The structure of the studieddispersions can be hypothesized as made of rheological units linked together by weak interactions (like the Vander Waals ones) that ensure stability to the structure. As a consequence, classical polymeric theories based onnetworks of strong covalent bonds are hardly extendable to these dispersions. Therefore, in order to correlate theviscoelastic behaviour with the structure of the dispersions, the data from both frequency and strain sweepmeasurements have been fitted according to the Gabriele et al. [26] and the Shih et al. [27] model, respectively.

The main characteristic of bentonite-isododecane dispersions is that, though being gel-like networks,they are capable of flowing because of a finite relaxation time. The structure of the material thus corresponds toa cooperative arrangement of flow units forming interactive strands. For such complex dispersions, the frequen-cy sweep data may be properly fitted with the weak gel model developed by Gabriele et al. [26] for foods andapplied to other kinds of dispersions [28]. According to this model, the complex modulus G* can be expressed by:

(1)

where AF is the gel strength, i.e. the strength of the interactions between flow units, while the coordinationparameter z represents the number of rheological units interacting with one another in the three-dimensionalnetwork. The fit parameters of Eq. 1 are reported in Table. 3.

With increasing the clay content, the gel strength (AF) always increases, confirming the G’ values. Thismeans that, independently of the number of interactions (z value), the system turns out to be stronger wheninorganic concentration increases. On the other hand, the coordination number z always decreases with the claycontent indicating a reduced number of interactions between rheological units but an enhancement of theirstrength due to the formation of larger aggregates. These latter are no longer able to move so easily at theimposed shear strain as the smaller aggregates, determining a loss of mobility that causes the decrease of thelinear viscoelastic region. Moreover, at a given inorganic content, the increase of calendering number promotesan increment of gel strength (AF) and a reduction of the coordination number z. This is likely due to a higherdegree of dispersion promoted by a better silicate delamination, which leads to a stronger reinforcement effecton the resulting nanocomposite. Therefore, the weak gel model of Gabriele et al. [26] provides a good explana-tion of the reduction of gcrit with the clay content and the number of calendering runs.

The results of strain sweep measurements have been quantitatively analysed by applying the modelproposed by Shih et al. [27] for predicting the rheological properties of colloidal gels containing dispersed tac-toids. The model considers the structure of a particle network as a collection of fractal aggregates or flocs thatare elastically linked together. It is possible to distinguish two different contributions to the elasticity of the gel:one comes from the elasticity of flocs, the second from the strength of interfloc links. Scaling of the elastic mod-ulus G’ and limit of linearity gcrit with increasing clay content allows the determination of the exponent from apower-law fit, which describes the fractal dimension. A decrease of the linearity limit with increasing volumefraction is associated with interfloc links stronger than the intrafloc links. The macroscopic elasticity of the gel isthus given by the intralinks [27, 29]. In the hypothesis of spherical colloidal gels, the storage modulus in the lin-ear viscoelastic region is predicted to have a power-law dependence on volume fraction f:

(2)

where df is the fractal dimension of the aggregates and x is the fractal dimension of the backbone relating the

G x d f' ∝ +( ) −( )f3 3

G G G AFz* ' '' /w w w w( ) = ( ) + ( ) =

2 2 1

To appear as:Appl.Rheol.19 (2009) 23423

Un

co

rre

cte

d P

ro

of

particle volume fraction with aggregate size. Likewise, the critical strain value can be related to the volume frac-tion as follows:

(3)

The linear regression of the plots in Figure 8 provide the power law exponents with which Eqs.2 and 3 are solvedsimultaneously to give the values reported in Table 4. It should be noted that, in order to obtain reliable plots inFig. 8, strain sweep measurements of dispersions with intermediate clay content (f = 6.5 %) have been carriedout.

The values of the extracted fractal dimensions (df) are comparable with those obtained for clay net-works in hydrophobic polymers such polypropylene or polystyrene [30]. The df values are in agreement also withthose obtained for clay networks in water and for colloidal gels made out of non-spherical colloids [31 - 32]. Thefractal dimension of aggregates formed by slow aggregation has been shown to be quite universal for variouscolloidal particles, ranging from colloidal golds to colloidal silica and to colloidal polystyrene particles [33]. Theeffect of shear application by calendering produce an increase in the backbone dimension x but a very slightdecrease in the fractal dimension df.

From the obtained results it appears that two factors contribute to the structural evolution of ben-tonite-isododecane dispersions: the orientation of the tactoids and the aggregation due to strong thermody-namic interactions. The attractive inter-particle forces create a more uniform particle network, whose elasticityand stability is enhanced by consecutive application of high shear forces as long as the clay content is below athreshold value. Above this critical value the further application of shear during processing of bentonite disper-sions accelerates the gel breakdown with a consequent decay of mechanical properties. The viscoelastic charac-terization here presented can be very helpful for quality control. In order to reach the desired gel elasticity, it canbe successful to use a powder content below the critical threshold and apply shear through subsequent calen-dering runs. The optimum combination of content and shear can be adapted to meet the required end-use per-formance.

5 CONCLUSIONSThe work has demonstrated the utility of applying rheological measurements for characterizing the viscoelas-tic behaviour of cosmetic nanofilled gels, obtained by dispersion of bentonite powder in isododecane. The test-ed samples have shown a gel-like behaviour with a final structure dependent on the clay content and shear his-tory applied by the processing method (mixing or mixing and calendering).The calendering produces a good degree of dispersion of the nanoclay in the solvent, as also confirmed by the X-Ray diffraction patterns. Nevertheless, the repeated calendering runs can cause the formation of particle-parti-cle and particle-solvent aggregates of different size and different structural level, thus accelerating the structur-al sample breakdown by further shear application. This phenomenon is strictly dependent on the clay content inthe dispersion as also demonstrated by the reduction of the critical strain. The dynamic mechanical results havebeen interpreted by using the models proposed by Gabriele et al. and the Shih et al.. The investigation present-ed in this paper, based on easy and fast measurements under small deformation conditions, can be used to deter-mine the proper clay content for nanofilled gel formulations in order to enhance their stability under manufac-turing and storage conditions.

ACKNOWLEDGEMENTSThe authors would like to kindly acknowledge Mr Donato Cannoletta of the University of Salento for the WAXSmeasurements. Mrs Sabrina Mottadelli, Mrs Laura Puglisi, Mr Giuseppe Maio of Intercos S.p.A. are also kindlyacknowledged for the helpful collaboration

REFERENCES[1] Viseras C, Meeten GH, Lopez-Galindo A: Pharmaceutical grade phyllosilicate dispersions: the influence of shear history

on floc structure, Int. J. Pharm. 182 (1999) 7-20.[2] Besün N, Peker S, Köktürk U, Yilmaz H: Structure of starch-bentonite gels, Colloid Polym. Sci. 275 (1997) 378-389.

g fcritx d f∝ − +( ) −( )1 3

To appear as:Appl.Rheol.19 (2009) 23423

Un

co

rre

cte

d P

ro

of

[3] Besün N, Ozgüclü B, Peker S: Shear-dependent rheological properties of starch/bentonite composite gels, Colloid Polym.Sci. 275 (1997) 567-579.

[4] Patel HA, Somani RS, Bajaj HC, Jasra RV: Nanoclays for polymer nanocomposites, paints, inks, greases and cosmetics for-mulations, drug delivery vehicle and waste water treatment, Bull. Mater. Sci. 29 (2006) 133-145.

[5] Tatum JP, Wright RC, US Patent 4752342 (1988).[6] Gadberry JF, Hoey M, Powell CF, US Patent 5663111 (1997).[7] Grim RE: Clay Mineralogy, McGraw-Hill, New York (1968).[8 ] Luckam PF, Rossi S:The colloidal and rheological properties of bentonite suspensions, Adv. Colloid Interface Sci. 82 (1999)

43-92.[9] Susteric Z, Kos T: Rheological Idiosyncrasies of Elastomer/Clay Nanocomposites, Appl. Rheol. 18 (2008) 54894.[10] Nsom B, Ravelo B, Ndong W: Flow regimes in viscous horizontal dam-break flow of clayous mud, Appl. Rheol. 18 (2008)

43577.[11] Bekkour K, Ern H, Scrivener O: Rheological Characterization of Bentonite Suspensions and Oil-In-Water Emulsions

Loaded with Bentonite, Appl. Rheol. 11 (2001) 178-187.[12] El-Sadi H, Esmail N: The Effect of Yellowing Inhibitor Total Charge on The Rheology of Paper Coating, Appl. Rheol. 12

(2002) 289.-296[13] Abend S, Lagaly G: Sol-gel transitions of sodium montmorillonite dispersions, App. Clay Sci. 16 (2000) 201-227.[14] Van Olphen H: An introduction to Colloidal Chemistry, Interscience, New York (1977).[15] Yao ML, Patel J: Rheological characterization of body lotions, Appl. Rheol. 11 (2001) 83-88.[16] Hamer G, Brummer R: Rheological methods to characterzie cosmetics, Appl. Rheol. 7 (1997) 19-24.[17] Bekkour K, Kherfellah N: Linear viscoelastic behaviour of bentonite-water suspensions, Appl. Rheol. 12 (2002) 234-240.[18] Bekkour K, Letama M, Benchabane A, Scrivener O: Time-dependent rheological behaviour of bentonite suspensions: An

experimental study, J. Rheol. 49 (2005) 1329-1345.[19] Alemdar A, Güngör N, Ece OI, Atici O: The rheological properties and characterization of bentonite dispersions in the

presence of non-ionic polymer PEG, J. Mat. Sci. 40 (2005) 171-177.[20] Güngör N: Effect of the adsorption of surfactants on the rheology of Na-bentonite slurries, J. Appl. Polym. Sci. 75 (2000)

107-110.[21] Galindo-Rosales FJ, Rubio-Hernandez FJ: Structural breakdown and build-up in bentonite dispersions, Appl. Clay Sci. 33

(2006) 109-115.[22] Mobuchon C, Carreau PJ, Heuzey MC: Effect of flow history on the structure of a non-polar polymer/clay nanocompos-

ite model system, Rheol. Acta 46 (2007) 1045-1056.[23] Lee JY, Lee HK: Characterization of organobentonite used for polymer nanocomposites, Mat. Chem. Phys. 85 (2004) 410-

415.[24] Wagener R, Reisinger TJG: A rheological method to compare the degree of exfoliation of nanocomposites, Polym. 44

(2003) 7513-7518.[25] Rueb CJ, Zukoski CF: Viscoelastic properties of colloidal gels, J. Rheol. 41 (1997) 197-218.[26] Gabriele D, De Cindio B, D’Antona P: A weak gel model for foods, Rheol. Acta 40 (2001) 120-127.[27] Shih WH, Shih WY, Kim SI, Liu J, Aksay IA: Scaling behavior of the elastic properties of colloidal gels, Phys. Rev. A 42 (1990)

4772-4779.[28] Martinez-Ruvalcaba A, Chornet E, Rodrigue D: Viscoelastic properties of dispersed chitosan/xanthan hydrogels,

Carbohydr. Polym. 67 (2007) 586-595.[29] Wu H, Morbidelli M: A model relating structure of colloidal gels to their elastic properties, Langmuir 17 (2001) 1030-1036.[30] Vermant J, Ceccia S, Dolgovskij MK, Maffettone PL, Macoscko CW: Quantifying dispersion of layered nanocomposites via

melt rheology, J. Rheol. 51 (2007) 429-450.[31] Pignon F, Magnin A, Biau JM, Cabane B, Lindner P, Diat O: Yield stress thixotropic clay supension: Investigation of struc-

ture by light, neutron, and x-ray scattering, Phys. Rev. E 56 (1997) 3281-3289.[32] Mohraz A, Moler DB, Ziff RM, Solomon MJ: Effect of monomer geometry on the fractal structure of colloidal rod aggre-

gates, Phys. Rev. Lett. 92 (2004) 155503-1- 155503-4.[33] Lin MY, Lidsay HM, Weitz DA, Ball RC, Klein R, Meakin P: Universality in colloid aggregation, Nature 339 (1989) 360-362.

To appear as:Appl.Rheol.19 (2009) 23423

Un

co

rre

cte

d P

ro

of

FIGURE CAPTIONS

Table 1: Studied samples as a function of the shear history and clay content.

Table 2: Storage modulus in the linear viscoelastic region and critical strain as a function of the shear history andclay content.

Table 3: Fit parameters according to Gabriele et al. model as a function of the shear history and clay content.

Table 4: Fit parameters according to Shih et al. model as a function of the shear history.

Figure 1: Diffraction spectra of the bentonite dispersions with 11 wt% clay compared with that of bentonite powder.

Figure 2: Diffraction spectra of the bentonite dispersions with 17 wt% clay compared with that of bentonite pow-der.

Figure 3: Effect of shear strain on the storage modulus of bentonite dispersions with different shear history and claycontent.

Figure 4: Effect of calendering runs on the viscoelastic moduli of bentonite dispersions with different clay content(‡: sample A, Ê: sample B). Filled and open symbols refer to G’ and G’’, respectively.

Figure 5: Effect of frequency on the storage modulus of bentonite dispersions with different shear history and claycontent.

Figure 6: Effect of consecutive frequency sweep runs on the storage modulus of bentonite dispersions (11 wt%) withdifferent shear history. Pre-shear treatment for the 2nd and 3rd frequency sweep run: stationary rotation at a shearrate of 20 s-1 for 10 minutes.

Figure 7: Effect of consecutive frequency sweep runs on the storage modulus of bentonite dispersions (17 wt%) withdifferent shear history. Pre-shear treatment for the 2nd and 3rd frequency sweep run: stationary rotation at a shearrate of 20 s-1 for 10 minutes.

Figure 8: Shih et al. model plots showing the storage modulus (filled symbols) and the critical strains (open symbols)as a function of clay volume fraction (‡: mixed samples, Ê: once calendered samples, Ú twice calendered samples).

To appear as:Appl.Rheol.19 (2009) 23423

Un

co

rre

cte

d P

ro

of

3

2

1

To appear as:Appl.Rheol.19 (2009) 23423

Un

co

rre

cte

d P

ro

of

6

5

4

To appear as:Appl.Rheol.19 (2009) 23423

Un

co

rre

cte

d P

ro

of

1

SAMPLE NAME

Shear history

Clay concentration 11 (wt %) 17 (wt %)

5 h stirred A-NC B-NC

Once calendered after 5 h stirring A-C1 B-C1

Twice calendered after 5 h stirring A-C2 B-C2

8

7

To appear as:Appl.Rheol.19 (2009) 23423

Un

co

rre

cte

d P

ro

of

4

x d f

N C ( n o t c a le n d e r e d ) 1 .4 3 2 .2 1

C 1 ( o n c e c a l e n d e r e d ) 1 .9 3 2 .2 0

C 2 ( tw ic e c a l e n d e r e d ) 2 .1 8 2 .1 9

3

Clay concentration

11% 17%

NC (not calendered) AF = 4.30*104 z = 61.8

AF = 2.36*105

z = 37.1

C1 (once calendered) AF = 8.02*104 z = 49.8

AF = 5.63*105 z = 22.4

C2 (twice calendered) AF = 3.80*105

z = 23.3 AF = 4.58*105

z = 22.8

2

C l a y c o n c e n t r a t io n

1 1 % 1 7 %

G ’ ( P a ) γ c r i t ( % ) G ’ ( P a ) γ c r i t ( % )

N C ( n o t c a le n d e r e d ) 3 .5 * 1 04

0 .9 5 .5 * 1 05

0 .2

C 1 ( o n c e c a le n d e r e d ) 6 .8 * 1 04

0 .5 1 .4 * 1 06

0 .1

C 2 ( tw ic e c a le n d e r e d ) 6 .5 * 1 05

0 .2 8 .0 * 1 05

0 .0 7

To appear as:Appl.Rheol.19 (2009) 23423