e/smectitehe elasticed.,e repulsive

lay particlee polymerinel is more

Journal of Colloid and Interface Science 273 (2004) 675–684www.elsevier.com/locate/jcis

Plastic and elastic properties of the systems interstratifiedclay–water–electrolyte–xanthan

O. M’bodj,a N. Kbir Ariguib,a M. Trabelsi Ayadi,b and A. Magninc,∗

a Laboratoire des Procédés Chimiques, Institut National de Recherche Scientifique et Technique, B.P. 95, 2050 Hamman-Lif, Tunisiab Faculté des Sciences de Bizerte, 7021 Zarzouna, Tunisia

c Laboratoire de Rhéologie, UJF Grenoble I, INPG, CNRS UMR 5520, B.P. 53, Grenoble Cedex 9, France

Received 22 July 2003; accepted 6 February 2004

Abstract

The present study concerns the rheological behavior of the Jebel Shemsi clay dispersions (Shp). Shp is an interstratified illitclay from southern Tunisia. The influences of clay concentration, NaCl, and xanthan—a semirigid polymer—on the yield stress, tmodulus, and the xanthan adsorption were investigated. The sol–gel transition and the scale laws of rheometric properties are establishProgressive addition of NaCl to the clay dispersions decreases the thickness of the diffuse double layer, which makes the system rigidincreasing the yield and the elastic modulus. In the presence of xanthan, the negative surface charges become higher and thinterparticle interactions increase; consequently the yield and the elastic modulus increase. The xanthan adsorption on the csurface increases slightly with the NaCl concentration. The particle aggregation due to the salt and the particle dispersion due to thare observed. The behavior of this interstratified clay is compared to that found for pure smectite. The 15% illite stratified with smectitethe Shp clay does not change the gels’ rheological properties significantly. Meanwhile the amount of Shp clay needed to obtain a gimportant than in the case of a pure smectite. 2004 Elsevier Inc. All rights reserved.

Keywords:Clay; Interstratified smectite–illite; Electrolyte; Polymer; Rheometry; Plasticity; Elasticity; Sol–gel

s of-ite–ro-

cro-pure

layThisatngress

ofo

, oracesi

m-tionstutedgo-.

highkeemf

ati-ofded

sarys inre

1. Introduction

The present study aims at investigating the influencean electrolyte additive (NaCl)and a semirigid polymer (xanthan) on the rheological behavior of interstratified smectillite clay dispersions. Relations between the clay macscopic behavior and the particle interactions at the miscopic scale will be deduced. Some comparisons with asmectite will be given.

The clay considered is provided from a bentonitic cdeposit located at Jebel Shemsi in southwest Tunisia.clay is formed by an interstratified smectite–illite [1] thmay be utilized in diverse industries, especially in drillifluids that need clay dispersions having a high yield stand certain stability over the time.

Smectites and illites are bidimensional phyllosilicatestype 2/1 [2]. The elementary layer is constituted from twtetrahedral layers of silicon oxide (SiO4)4− on the two sides

* Corresponding author.E-mail address:magnin@ujf-grenoble.fr (A. Magnin).

0021-9797/$ – see front matter 2004 Elsevier Inc. All rights reserved.doi:10.1016/j.jcis.2004.02.028

of an octahedral layer formed by aluminum, magnesiumiron oxide. The particles have negative charges on their fdue to isomorphic substitutions. In the tetrahedral sites S4+can be replaced by Al3+, and in the octahedral sites Mg2+can replace Al3+. The negative charges, in excess, are copensated by interlayer cations. In the smectites these caare simply inserted between the layers and can be substiand solvated, while in the illite they are inserted in hexanal cavities and are neither exchangeable nor solvatable

The smectites have a high specific surface and aaffinity for the water and other polar solvents, which mathem swell. They also have flexible particles that give thplastic and elastic properties [3]. The illites have none othese properties.

It is assumed that the presence of illite in the interstrfied clays affects the flexibility and the swelling propertiesthese clays; thus it affects the rheological properties neefor drilling fluid clay.

To master the macroproperties of a clay it is necesto understand the clay behavior on the microscale. Athe clay–water system, the clay flexibility and swelling a

676 O. M’bodj et al. / Journal of Colloid and Interface Science 273 (2004) 675–684

thesaryits

ater,sheetionthe

en-ted

Sev-f thesur-et haees

thehee toou-

the

,

ainthe

yer[8]ter-

ly totrac

reedgepro-

thentac

Onmedged,ivelythorncet ge

yerre-pH

luted) bye-con-

y ofre-layerhy-

ture:eat-pic

tsthe

EFet al.rmed

hasclay

nclay.on-ad-sed

eenonofuch

tedthesitythes-n,ibuteto ao-clescen-as

ff the

ures [5,

ise ad-bed

of

influenced by the clay fraction, the ionic strength, andadditives. To obtain the clay micro properties, it is necesto determine the clay rheological characteristics, such asyield stress and its elastic modulus in different conditions.

When the smectite particles are in suspension in wthe hydration of the compensator cations present on thesurface provokes a decrease of the electrostatic attracTherefore these cations have the tendency to diffuse inmiddle and to form an electrical double layer. The compsator cations then form a diffused cloud, thermally agitabut globally kept by a charged surface of opposed sign.eral theories have been proposed for the description odistribution of the ions that are adjacent to the chargedfaces in colloids. The Gouy–Chapman theory of the diffusdouble layer has received the greatest attention, and ibeen applied to the behavior of clays with varying degrof success [4,5].

According to the Poisson and Boltzmann equations,study of the electrical potential evolution that follows tnormal direction of the particle surface makes it possiblobtain the two principal characteristic parameters of the dble layer:

— the electrical potential on the shear plane, calledζ po-tential and denotedζ ;

— thek−1 term, called the Debye length, representingthickness of the double layer,

(1)1

k=

√εε0RT

2F 2I,

whereF is the Faraday constant (96.485 C),ε is thedielectric permittivity,ε0 is the permittivity of vacuumand I = (1/2)

∑ν2i Ci is the ionic strength (νi,Ci are

valence and concentration of the ioni).

Luckham and Rossi [5] and Abend and Lagaly [6] explthat the thickness of the diffuse layer depends only onionic strength (at constantε andT ).

From the knowledge of the structure of the double laDerjaguin and Landau [7], and Verwey and Overbeekhave developed the DLVO theory to understand the inactions between the particles. This theory leads globalrepulsive electrostatic forces at long distances and to attive forces (van der Waals) at short distances.

When the clay particles are in the suspension, thmodes of association are possible; face-to-face (FF), eto-edge (EE), and edge-to-face (EF). The first modelposed to describe the mode of particle association insuspensions where the particles begin to come in cohas been the card-house structure of Van Olphen [9].the basis of rheological observations Van Olphen assuthat, at acidic pH, the particle edges are positively charwhich leads to EF contacts with the faces that are negatcharged. This idea has been developed by several au[10,11]. But other studies have led to denial of the existeof such types of assemblages [12–15]. They assume tha

t.

s

-

-

t

s

l

formation is due to the long-range electrostatic double-larepulsion (EE and EF). Benna et al. [16] think that thesesults are not contradictory and that all depends on theconsidered.

On the other hand, the observations made on semidisuspensions of smectite at natural pH (isoelectric pointMET [17,18] completed by DPAX [19] show a heterogneous system formed by isolated sheets and particlestaining a variable number of parallel sheets. The studthe distribution of the interlayer distances in the particlesveals distances greater than 4 nm, related to the double-repulsion, and others less than 2 nm, corresponding todrated states.

On a larger scale, the SEM reveals an alveolar structhe system is formed by lenticular pores for which the grest dimension is of micrometer order. These microscostudies do not confirm any association of EF type.

Van Damme [20] thinks that the flexibility of the sheeand of their packets are the essential elements that allowsystem to realize the lenticular pore structure, withoutassociations. To integrate this component, Van Dammepropose a model based on a connected lenticular net foby a random aggregation of sheets. This type of structurebeen observed several times by other authors [21,22] onfiltration cakes in the presence of salt and polymers.

To improve the drilling fluids’ rheological and filtratioproperties, different products are usually, added to theThese products interact with the electrical double layer. Csequently they influence the interparticle interactions, leing to a change in the clay macro behavior. The most uadditives are electrolytes and the polymers.

In the literature, NaCl as an electrolyte additive has bthe most considered. The influence of its concentrationrheological properties, viscosity, and elastic propertiespure smectite (and pure kaolinite) dispersions has been minvestigated [6,14,23,24]. Some works carried out on diluand semidiluted suspensions of smectite show that, inpresence of NaCl, the yield stress and the plastic viscopresent a minimum for a low NaCl concentration. WhenNaCl concentration increasesthe yield stress and the platic viscosity grow. For relatively high NaCl concentratiothe rheological parameters decrease. The authors attrthe minimum of the yield stress and the plastic viscositydiminution of viscoelastic effects. The increase of the rhelogical parameters is due to the flocculation of the partiby EF and EE contacts. They thus consider the salt contration corresponding to the beginning of this flocculationthe critical concentration of flocculationCk . The decrease othe rheological parameters is attributed to the collapse ointerparticle network.

The interactions of polymers with pure smectite (and pkaolinite) dispersions are also treated in several studie21,25–27] that show that when a polymer in solutionadded to a clay suspension, the interactions lead to thsorption of the polymer by the clay. The polymer adsoronto the surface of the clay particles appears in the form

O. M’bodj et al. / Journal of Colloid and Interface Science 273 (2004) 675–684 677

oly-elyh deth),, ththere-es,tro-t, itof

clayn orivee isci-hecanrn-ible

st-to

tudyin-

eldtions

estsi-

car-age

es-s a%

argescity

ite

dhp

, a-

aure-rallyionsnm0 tormstive

raluc-dper-.ble.red

atdato-

eana

s an-nt is

tro-ilitym-

d inn it

inh,

d toen

is-n at

at,inedAtwa-fterer–lokh.

head, train and tail. This adsorption depends on the pmer molecule dimensions and flexibility, which are closrelated to the persistence length. This persistence lengtpends on the molecular chemical structure (intrinsic lengthe repulsions between the charged groups on the chaincharge density of the polymer, and the ionic strength ofmiddle. The interactions during the adsorption are thesult of the competition between the van der Waals forchydrogen bonding, and the attractive or repulsive elecstatic forces. When a salt is added, by its screen effecfacilitates polyanion polymer adsorption on the surfacethe clay particles. The polymer layer adsorbed onto theparticle surface can lead to the dispersions’ stabilizatioto their flocculation. The stabilization due to the repulseffect of macromolecules present on the particle surfaccalled steric stabilization [5,28]. The flocculation is assoated with the formation of interparticle bridges or with tinterpenetration of the adsorbed polymer layers, but itbe by depletion [28,29]. The majority of the studies conceing the clay–polymer–water system are related to the flexpolymers and sometimes to the semirigid ones [21,27].

As xanthan is a semirigid polymer that can give intereing properties to the drilling fluids and as NaCl seemsinfluence the clay rheological properties, the present sdeals with the plastic and elastic properties of the systemterstratified smectite/illite–NaCl–xanthan–water. The yistress and the elastic modulus will be measured as funcof clay, NaCl, and/or xanthan concentrations.

2. Materials and methods

2.1. Starting materials

The clay studied comes from Jebel Shemsi (southwTunisia). Shb symbolizes it. After purification by a clascal method [16] it is denoted by Shp.

The complete characterization of Shb and Shp wasried out previously [1]. It has been shown that the averstructure formula of Shp is

(Si7.585Al0.415)(Al2.757Mg0.475Fe0.725Ti0.094Mn0.0014)

(Na0.449Ca0.026K0.088).

The crude Shb contains 81% clay and 19% impurities,sentially quartz and calcite [1]. The clay fraction Shp idioctahedral interstratified smectite–illite, containing 80smectite, 15% illite and some free kaolinite (≈5%). Thetetrahedral charge deficit is 0.415 and the octahedral chdeficit 0.220; this indicates that the smectite fraction haa beidellitic character. The Shp cation-exchange capa(CEC) is 87 meq/100 g of calcined clay, a pure smectCEC ranges between 80 and 150 meq/100 g [30]. The Shpspecific surface area is 711 m2/g; it ranges between 700 an800 m2/g for a pure smectite [31]. The density of the Sclay is 2500 kg/m3.

-

e

The water used is deionized water.The sodium chloride used has purity equaling 98%.The polymer used in the present study is xanthan

biopolymer obtained by fermentation of theCampestris xanthomonasbacterium. Its double helix structure gives itgreat rigidity over a large range of salinity and temperat[32]. Its chemical structure is a principal chain formed by Dglucose molecules on which a trisaccharide chain is latefixed each two glucose molecules. The average dimensin ordered form are 800 to 1200 nm for the length and 2for the diameter. The persistence length ranges from 5120 nm. The xanthan structure studies show that two foare conceivable [21]. In the presence of salt, the negacharges present in the lateral chains are screened. The latechains are folded away on the principal chain. The strture is a simple or double helixaccording to the salinity antemperature conditions. In absence of salt or at high temature, the xanthan is dissociated in flexible simple chainsThis structure is much less chemically and thermally staThe xanthan used is the RHODOPOL 23. It is manufactuby Rhone Poulenc. Its molecular mass is MW= 2× 106.

The polymer aqueous solution is shaken at 350 rpm50◦C for 2 h in a type RAYNERI shaker. The obtainesolution is filtered at 21◦C under a pressure of 0.15 MPthrough a filter with a pore diameter is 1.2 µm, aimingeliminate the microgels and to obtain a transparent solution. The xanthan solution viscosity is measured by mof an ARES rheometer from Rheometric Scientific, withcone–plane geometry. The cone diameter is 5 cm and itgle is 0.04 rad. The temperature during the measureme21± 0.5 ◦C.

2.2. Dispersions preparation

The preparation mode of the mixtures (water/clay/eleclyte/polymer) has a considerable influence on their stabover time and their rheological behavior [21]. All the saples are prepared according to the following methods:

Dispersions of clay in water: the clay Shp is dispersewater. The mixture is shaken at 180 rpm for 4 h and therests for 48 h.

Dispersions of clay–water–salt: the clay is disperseddeionized water, the mixture is shaken at 180 rpm for 4and then it rests for 24 h. Then the salt solution is addethe clay dispersion, which is mixed again for 4 h and thleft for 48 h.

Dispersions of clay–water–salt–polymer: the clay is dpersed in the deionized water and the mixture is shake180 rpm for 4 h and then left resting for 20 h. After ththe salt solution is added in the clay dispersion; the obtamixture is shaken for 4 h and then left resting for 20 h.the same time the polymer is dissolved in the deionizedter. The salt solution is added to the polymer solution. Athat the salt–polymer solution is mixed with the clay–watsalt dispersion. The mixture is intensely mixed by a Biobmixer for 30 min and then by a shaker at 220 rpm for 4

678 O. M’bodj et al. / Journal of Colloid and Interface Science 273 (2004) 675–684

theidforewer

r-

080andand

inedthe

d-the

ingsamingthatarge

theTheetering

foper-3,35-etere).d byandonals aense.sticre-

sionandThepec

t

oids a

thehavefn in

tingd by

ge:

sys-on-andresslay

ri-willism

ion.willf theThisad-

od-t athethe

32lue

The obtained dispersion is left resting for 48 h. Beforemeasurements no preshear is applied to the samples, to avoany modification of the structure of the suspensions bethe yield stress and the elastic modulus measurementsreached in the virtual staticconditions. During the expeiments the temperature is maintained at 21◦C and the pHfrom 7 to 8. The clay volume fractions range between 0.and 0.12, the NaCl concentrations are between 0.0010.02, and the polymer concentrations are 500, 1000,2000 ppm.

2.3. Rheometry

The yield stress and the elastic modulus are determby the vane method. The use of the vane instead ofmore conventional measuring geometries has two major avantages [33]. First, any wall slippage is absent sinceyield surface is within the material itself. Secondly, insertthe vane causes much less structural disruption to aple than introducing the fluids into conventional measurgeometries. This is important for fluids, such as ours,have a fragile gel structure which can be destroyed by lstrains. This method avoids perturbing the sample duringmeasurement and its slipping on the tools’ sides [34].vane used is formed by four blades. The blades’ diamis 15 mm and their length is 10 mm. The bottles containthe samples measure 4 cm for the diameter and 4.5 cmthe length. The usefulness of the vane method in clay dission studies has been demonstrated several times [16,3The values of the yield stress (σ ) and of the elastic modulus (G) are determined by a Weissenberg Rheogoniomtype CARRI-MED rheometer at a controlled rate (CR-typThe yield stress and the elastic modulus are measureimmersing the blades of a vane slowly in the samplethen making the blades rotate slowly at constant rotatispeed (10−3 rad s−1). The resulting torque is measured afunction of the time. A typical torque–time curve is givin Fig. 1. The region A shows an initial transient responThe linear behavior in region B corresponds to the elaresponse of the sample. From the slope of the curve ingion B, the elastic modulus of the suspensionG can beevaluated. The maximum torqueTm given by region D of thetorque–time curve allows the yield stress or the suspenσ , to be evaluated once the geometry of the yield surfacethe shear stress distribution on this surface are known.yield stress and the elastic modulus are determined restively according to the relations

(2)σ = Tm/(πD3/2)(H/D + 1/3),

(3)G = (1/4πωH)(dT/dt)(1/R2 − 1/R2

c

),

whereTm is maximum torque,T is torque,D is diameterof the blades,H is height of the blades,R is radius of theblades,Rc is radius of the bottle containing the sample,ω isrotation speed of the blades, andt is time. Details of themethod of calculatingσ and G are given by Alderman eal. [33].

e

-

r

].

,

-

Fig. 1. Schematic torque–time plot for a vane test.

The blades are gently introduced into the fluid to avits perturbation. The variation of the torque is plotted afunction of the time.

2.4. Dosage of carbon

The dispersions containing the xanthan and for whichyield stress and the elastic modulus were measuredbeen centrifuged at 17,000 t/min for 2 h. The quantity opolymer adsorbed was determined by dosing the carbothe supernatant and in the initial solution and calculathe difference between them. The carbon was measuremean of a DOHRMANN DC-190 carbon analyzer (ran0.2–1000 ppm carbon with±2%).

3. Results and discussion

The rheometric measurements carried out on thetem interstratified smectite/illite–NaCl–xanthan–water ccern the determination of the Shp rheological properties,otherwise the plastic and elastic properties. The yield stand the elastic modulus variation as a function of the cfraction will define the point of sol–gel transition; their vaation as a function of NaCl and xanthan concentrationsallow a better understanding of the interaction mechanof different additives an interstratified smectite dispersMoreover, the polymer adsorption on the clay surfacebe measured by carbon analysis after centrifugation odispersions, which achieved rheometric measurements.analysis gives the exact amount of the polymer reallysorbed onto the clay surface.

3.1. Effect of the clay fraction on the yield stress and theelastic modulus

The measurement of the yield stress and the elastic mulus as functions of the clay volume fraction is carried oupH 8.1, which is the natural pH of the clay Shp in water. Tresults Fig. 2 show that the yield stress increases withclay volume fraction. At a volume fraction equal to 0.0(8% w/w) the value of the yield stress is 18.1 Pa. This va

O. M’bodj et al. / Journal of Colloid and Interface Science 273 (2004) 675–684 679

rac-

ess

ivenly

.erel6].dandtheol-the

clay

-byen-n-

tionases

-

ticle

lus

ressen-

aseaClum

ofationrcesse ofpro-weend forsedi-n be-ent

se of

ivess.ains

Fig. 2. Yield stress and elastic modulus variation as functions of clay ftion.

corresponds to the first clay volume at which a yield strdifferent from zero is detected. As the clay volume fractiontested just below 0.032 is 0.028 (7% w/w) and does not gany yield value, the sol–gel transition of Shp clay certaihappens between 0.028 and 0.032. The yield stress (σ ) fol-lows the scale law,σ = A(φv)

B [13], where (φv) is the clayvolume fraction. The parametersA andB determined fromour results (Fig. 1) are respectively 1.2 × 1011 Pa and 6.5The parametersA and B for the Wyoming pure smectit(Wp), defined under the same conditions as those of Shp, arespectivelyA = 3 × 1011 Pa andB = 6.1 and the sol–getransition occurs for a volume fraction equal to 0.022 [3A comparison between theA andB parameters of Shp anWp and the volume clay fraction at the sol–gel transitionour results shows that theA parameter decreases whenillite is present but theB parameters are close and the vume clay fraction at sol–gel transition increases undersame conditions.

Moreover, according to Ramos et al. [37], theB parame-ter for the stable dispersions ranges between 5 and 8.

The elastic modulus measured as a function of thevolume fraction Fig. 2 follows the power lawG = CφD ,where C and D values are respectively 8.48 × 108 Paand 4.3.

The variation of the yield stress (σ ) and the elastic modulus (G) according to the scale laws can be explainedthe fact that the gel is formed by an interconnected tridimsional network due to the flexibility and the big lateral extesion of the smectite sheets. When the clay volume fracincreases the number of particles per volume unit incre

Fig. 3. Yield stress and elastic modulus variation as functions of NaCl concentration.

and this leads to an increase of the number of interparinteractions.

3.2. Effect of NaCl on the yield stress and the elasticmodulus of the system Shp–Water

The variation of the yield stress and the elastic moduof Shp dispersion [8.5% w/w (φv = 0.034)] as a functionof NaCl concentrations (0 to 0.381 mol L−1, 0 to 2% w/w)(Fig. 3), shows that in the presence of NaCl the yield stand the elastic modulus increase rapidly with the conctration. They reach a maximum at 0.112 mol L−1 of NaCl(0.6% w/w). However, it is to be noted that the decreof the yield stress after the maximum is sudden. The Nconcentration at the maximum is attributed to the maximconcentration coagulation (MCC).

For relatively small NaCl concentrations, the increasethe yield stress can be explained by progressive aggregof the clay particles due to a decrease of repulsion focaused by electrical double-layer compression (decreathe Debye length) due to a partial screening effect. Thisgressive aggregation allows an increase of contacts betthe particles. However, as the dispersions that had servethe tests are stable and do not show any macroscopicmentation, the increase of contact between particles caaccompanied by an eventual reorganization of the tridimensional network. The system moves with a rapprochembetween the particles in the aggregates and an increathe macropore volume.

For the higher NaCl concentration, the aggregation grigid particle packets. These packets form oriented domainHowever, the orientational order of the aggregates rem

680 O. M’bodj et al. / Journal of Colloid and Interface Science 273 (2004) 675–684

. Ins, buf Elsoateshustheand

12ach

rsiobe-tingctiv

ticleands,

-ate

heyiel

aveviorfor

as-ion.ns

ingoumlayrly

be-o

CC,C,nn-rigid

ameintma-

on-casep-

and

thetresslowtionheble

tion.sentstedtra-ssberof

os-

ons,ofarin-

rved.Cl

pm)for

theoesa-as

local and is not extended to the entire sample volumethis structure FF contacts are present in the aggregatebetween the aggregates the contacts are essentially otype [5]. As the aggregates are sufficiently thick they acan form EE contacts [24] (see Fig. 10B). In the aggregthe particles are bound by the van der Waals forces. Tthe rigidity and the random aggregate orientation makeshear of the suspension difficult, so that the yield stressthe elastic modulus increase again.

For NaCl concentrations higher than the MCC (0.1mol L−1), the yield stress and the elastic modulus rea maximum, which means that the forces maintainingthe three-dimensional network decrease and the dispeneeds no more constraint to flow. The screening effectcomes total when the van der Waals attractive forces acbetween the particles become predominant. These attraforces increase the aggregation degree of the clay parin the dispersion and lead to the formation of compactrigid aggregates. The total screening between aggregatetheir rigidity, and the decrease in the effective volume fraction act to break down a significant number of interaggregcontacts. The majority of aggregates are thus isolated and tsystem flocculates, leading to a sudden decrease of thestress.

The dispersions at 0.018 and 0.131 mol L−1 of NaCl,which are situated on the two sides of the MCC, hthe same value of yield stress, while their elastic behais different. Indeed, the value of the elastic modulus0.131 mol L−1 of NaCl is greater than that for 0.018 mol L−1

of NaCl. According to Eq. (3), one can recall that the eltic behavior is related to the strain kinetics of the dispersThe variation of the torque of the two different dispersio(NaCl concentrations: 0.131 and 0.018 mol L−1) is plottedas a function of the time (Fig. 4). The rotation speed beconstant and very low (10−3 s−1), the time corresponds ta strain applied to dispersion. The strain at the maximtorque corresponds to the rupture of the network of cparticles and to the flow of dispersion. This shows cleathat the strain when NaCl concentration is 0.131 mol L−1 issmaller than that when NaCl concentration is 0.018 mol L−1.This observation can be explained in terms of contactstween the particles and their rigidity or flexibility: as the twNaCl concentrations are ranged on both sides of the Mthe textures of the two gels are different. Before the MCthe tridimensional networkis formed by contacts betweeflexible particles, while after the MCC, the number of cotacts between the particles decreases and they becomeConsequently, even if the two dispersions exhibit the svalue of yield stress, the origin of the level of this constrais not the same in the two cases, which is why their defortions are different. Indeed, in the first case (0.018 mol L−1 ofNaCl), the origin of yield stress is the breakdown of the ctacts between the flexible particles, and in the second(0.131 mol L−1 of NaCl), the origin is mainly due to the searation of weakly connected rigid particles.

tF

n

es

d

.

Fig. 4. Torque variation as a function of time for suspensions at 0.6834.781 mmol L−1 of NaCl.

Previous studies concerning the effect of NaCl onyield stress or apparent viscosity or extrapolated shear svariation have shown the presence of a minimum forNaCl concentrations [5,13,24,38]. They used a clay fraclower than 5.5% w/w. They explained this minimum by tpossibility of the particles moving, as the area of the doulayer is large, in the presence of a weak NaCl concentraThe absence of the minimum of yield stress in the prestudy can be attributed to the fact that the dispersions teare gels (8.5% w/w of clay), and at weak NaCl concentions (0.018 mol L−1 of NaCl) Shp gel exhibits a yield strevalue of 128 Pa. Moreover, in a gel structure, the numof particles per volume unit leads to an interconnectionparticles in the three-dimensional network, limiting the psibility of the particles moving.

3.3. Effect of NaCl on the system xanthan–water

The viscosity variation of xanthan aqueous dispersiat different concentrations, was examined as a functionshear rate (Fig. 5). The viscosity decreased with the sherate (shear thinning). However, the viscosity of xanthancreased with its concentration. No yield stress was obse

The viscosity variation of xanthan as a function of Naconcentration (8.55× 10−3 to 8.59× 10−3 mol L−1, 0.05 to0.5% w/w) for a constant xanthan concentration (2000 pis studied. Fig. 6 shows that the viscosity decreaseslow NaCl concentration (8.55 × 10−3 mol L−1) and in-creases slightly with NaCl concentration. Nevertheless,increase of the viscosity with the NaCl concentration dnot reach the level of the viscosity of xanthan in the wter. Indeed, for low shear rates, the xanthan viscosity

O. M’bodj et al. / Journal of Colloid and Interface Science 273 (2004) 675–684 681

ntra-

umeaterth,,

ions.

is-oonicase

n innat-on-

that

lumens

tions

sur-,withndsalt,,25,andic,

rp-thee of

Fig. 5. Viscosity of xanthan as a function of the shear rate.

Fig. 6. Viscosity of xanthan as a function of shear rate and NaCl concetion.

a function of NaCl concentration presents a low minimat 8.55 × 10−3 mol L−1. Many works carried out on thsecondary structure of xanthan show that xanthan in wpresents a double helix structure. At very low ionic strengthe xanthan can be decomposedinto simple chains [21,39

Fig. 7. Adsorbed carbon as a function of xanthan and NaCl concentrat

40]. This denaturation can explain the diminution of the vcosity at 8.55× 10−3 mol L−1 NaCl. The same works alsshow a renaturation of the xanthan molecule when the istrength increases. That could explain the viscosity increwith NaCl concentration (�1.71× 10−2 mol L−1 NaCl). Atthe same time the viscosity does not reach that of xanthapure water. This result can be attributed to incomplete reuration and decreased xanthan rigidity. Indeed, Camesanand Wilkinson [40] have studied the evolution of the xathan structure in the presence of KCl; they also showthe xanthan rigidity decreases in the presence of salt.

3.4. Effect of NaCl on the system Shp–xanthan–water

The dispersions used as samples contain a clay vofraction equal to 8.5% (w/w). The polymer concentratioare 500, 1000, and 2000 ppm and the NaCl concentraare 0.018 and 0.093 mol L−1.

3.4.1. AdsorptionThe xanthan amount adsorbed onto the clay particle

faces as a function of its concentration is reported in Fig. 7which shows that at 0% NaCl the adsorption increasesxanthan concentration, reaching a plateau between 1000 a2000 ppm. The xanthan adsorption, in the absence ofoccurs by van der Waals forces and hydrogen bonding [541]. The hydrogen bonds are built between the hydrogenthe hydroxyl of clay and those of the groups D-glucoronD-mannose, and pyruvate of the xanthan molecule.

In addition, at high electrolyte concentrations, the adsotion of polyelectrolytes with the same charge sign onparticle surface is enhanced [29]. Indeed, in the presenc

682 O. M’bodj et al. / Journal of Colloid and Interface Science 273 (2004) 675–684

s.

iss ofd

d

s ofions

in-a-ads

ginge canceula-ameility,per-insicona

r onan-the

t byter-ctsad-artic

ons.

therel-

be-the

con-and

truc-s are

hanetheno-

e oftress

the

sid-

e

Loc-0C)

Fig. 8. Yield stress as a function of xanthan and NaCl concentration

NaCl, the adsorption of xanthanincreases. This increaseprobably due to the adsorption of xanthan on the faceclay particles, because as the xanthan is a negatively chargemacromolecule, in the presenceof electrolyte it is adsorbeby a screen effect [21,25,29].

3.4.2. Yield stress and elastic modulusThe yield stress and the elastic modulus as function

xanthan concentration and at different NaCl concentratare reported in Figs. 8 and 9.

At 0% NaCl the yield stress and the elastic moduluscrease with the xanthan concentration. In general, whenmacromolecule is adsorbed onto a colloid surface, it interconnects several particles. This interparticle bridging leto an increase of the viscoelastic properties. The bridphenomenon occurs when the adsorbed macromoleculextend far enough from the particle surface to a distawhere the electrical repulsion operates [5,42]. The flocction of the colloidal suspensions depends on several parters, such as the polymer molecule dimension and flexibwhich are closely related to the persistence length. Thissistence length depends on the chemical structure (intrlength) of the molecule. However, the recent studies donethe xanthan molecule confirmthat the native xanthan hasdouble-helical structure, which confers a rigid characteit [40]. In the last reference it is indicated that when the xthan is in pure water it is adsorbed flatly on the surface ofmica. The results of the studies of Loeber [22] carried ouSEM and TEM reveal that the xanthan does not form inparticle bridging. It is flatly adsorbed and it interconneat most two clay particles. Thus, due to its ionicity, thesorbed xanthan increases the negative charges on the p

n

-

le

Fig. 9. Elastic modulus as a function of xanthan and NaCl concentrati

surfaces. This increases the level of repulsion betweenclay particles [21,27]. The increase of the repulsions in aatively large effective volume fraction (Shp gel,φv = 3.4%,8.5% w/w) (Fig. 10B) reinforces the number of contactstween the particles; consequently the yield stress andelastic modulus increase.

In the presence of 0.018 mol L−1 of NaCl, the yieldstress and the elastic modulus increase with the xanthancentration and reach a pseudo-plateau between 10002000 ppm. In this case, on one hand the suspension sture remains close to that observed when the suspensionwithout salt. In that range the quantity of adsorbed xantincreases with its concentration (Fig. 7), which increases thnumber of interparticle repulsions. On the other hand,double layer is lightly compressed and does not changeticeably the effective volume fraction. Thus, the increasinteractions number leads to the increase of the yield sand the elastic modulus.

In presence of 0.093 mol L−1 of NaCl, the curve is in-versed, the yield and the elastic modulus decrease withxanthan concentration.

To understand this behavior two effects must be conered:

(1) The NaCl effect on the clay.(2) The NaCl effect on the xanthan and its adsorption on th

clay particles.

In the presence of a high NaCl concentration (0.093 mol−1

of NaCl) the suspension structure of Shp is strongly flculated, the particles forming rigid aggregates (Fig. 1

O. M’bodj et al. / Journal of Colloid and Interface Science 273 (2004) 675–684 683

.5%

ofiblenot

ad-. In-the

re-on-min-eenn adm anre ishethe

dulu

at-tsme

ing

ioninggel

w-

ing

tra-the

w of.1%pm

thethethe

-in-

er-ve),in-las-he

r bydis-

to

e-olo-anro-

n

Fig. 10. Schematic systems: (A) Shp (8.5% w/w)–water; (B) Shp (8w/w)–water–xanthan; (C) Shp (8.5% w/w)–water–NaCl (0.5% w/w);(D) Shp (8.5% w/w)–water–NaCl (0.5% w/w)–xanthan.

(Section 3.2). This structure is quite different from thatweak salinity suspensions where the particles are flex(Fig. 10A). The xanthan effect on the two structures isthe same.

When the NaCl concentration increases, the xanthansorption increases, but its ability to disperse decreasesdeed, at high ionic force, the charge screen effect givespolyelectrolytes a more compact configuration [29]. Moover, Zhou et al. [28] show that when the electrolyte ccentration increases, the double-layer compression diishes the effective volume fraction. The repulsion betwthe aggregates leads to their isolation. Thus, the xanthasorbed on the aggregates prevents contact between thethe tridimensional network becomes discontinuous. Theelectrosteric stabilization of the suspension. Therefore, tnetwork collapse and the isolation of the particles makeshear easier so that the yield stress and the elastic modecrease.

-d

s

4. Conclusion

The Shemsi purified clay fraction (Shp) is an interstrified smectite–illite containing 15% illite. The study of iyield stress and its elastic modulus as functions of its volufraction and the NaCl concentration leads to the followconclusions:

The Shp sol–gel transition occurs at a volume fractequal to 0.032. This value is higher than that of the Wyomsmectite, equal to 0.022. It is, then, clear that the sol–transition depends on the presence of illite.

After the sol–gel transition, the yield stress varies folloing the scale lawσ = A(φv)

B . The B value found for theShp clay are in the same range as those of the Wyompure smectite.

The study of the effect of xanthan and NaCl concentions on Shp gels makes it possible to conclude that forweak and mean aggregated dispersions (0 to 0.1% w/NaCl) the xanthan has a dispersant effect, which, at 0w/w of NaCl, remains constant between 1000 and 2000 pof xanthan.

For the aggregated suspensions (0.5% w/w of NaCl),decrease of the effective volume fraction of particles anddisperse effect of the xanthan lead to the diminution toyield stress and the elastic modulus.

The preceding results make itpossible to relate the formation of a stable gel in an aqueous dispersion of anterstratified smectite/illite to the nature of the particle intactions, which are either electrostatic (repulsive/attractivan der Waals, or mechanical (flexibility/rigidity). Theseteractions governing the magnitude of elasticity and pticity are a function of the clay amount as well as of telectrolyte and the polymer concentrations.

Therefore, the NaCl coagulates the system Shp-wateits screening effect and the xanthan stabilizes it by itspersing effect.

Meanwhile, to obtain a good suspension for the drillingfluids, it is necessary to optimize the additive amountscompensate for the smectite deficit in the Shp.

Acknowledgments

We are grateful for financial support provided by the “Scrétariat d’Etat à la Recherche Scientifique et à la Techngie” of the Tunisian government to the project “Tunisiclays and drilling muds.” The authors are grateful to Pfessor Jean Michel Piau for his help.

References

[1] O. M’bodj, M. Ayadi, M. Benna, N.K. Ariguib, M.T. Ayadi, Asian J.Chem. 15 (3) (2003).

[2] S. Caillere, S. Henin, M. Rautureau, Minéralogie des argiles, EditioSeptima, Paris, 1989.

684 O. M’bodj et al. / Journal of Colloid and Interface Science 273 (2004) 675–684

aisestière

6,

92.

)

ho-

,

22–

974)

ara-

57–

of992,

e

Int.

ol-91,

.nal92–

, F.ed

desr-

le,

.

face

t. 52

01–

mp-o-

5–

id

–

)

-de

ace

iol.

191.Inst.

[3] H. Van Damme, Physique et mecanique des boues et des glCNRS et Université d’Orléans, Centre de Recherche sur la MaDivisée, 1994.

[4] J.K. Mitchell, Fundamentals of Soil Behavior, Wiley, New York, 197pp. 31–78.

[5] P.F. Luckham, S. Rossi, Adv. Colloid Interface Sci. 82 (1999) 43–[6] S. Abend, G. Lagaly, Appl. Clay Sci. 16 (2000) 201–227.[7] B.V. Derjaguin, L.D. Landau,Acta Physicochim. USSR 14 (1941

633.[8] E.J.W. Verwey, J.Th.G. Overbeek, Theory of the Stability of Lyop

bic Colloids, Elsevier, Amsterdam, 1948, p. 216.[9] H. Van Olphen, Introduction to Clay Colloid Chemistry, Interscience

New York, 1963.[10] R.K. Khandal, T.F. Tadros, J. Colloid Interface Sci. 125 (1988) 1

128.[11] R. Shom, T.F. Tadros, J. Colloid Interface Sci. 132 (1989) 62–71.[12] I.C. Callaghan, R.H. Ottewill, Faraday Discuss. Chem. Soc. 57 (1

110.[13] B. Rand, E. Pekenc, J.W. Goodwin, R.W. Smith, J. Chem. Soc. F

day Trans. 76 (1980) 225–235.[14] J.S. Chen, J.H. Cushman, P.F. Low, Clays Clay Miner. 38 (1990)

62.[15] M. Moan, in: H. Toulhat, J. Leeovrtier (Eds.), Physical Chemistry

Colloids Interfaces and Oil Production, Editions Technip, Paris, 1pp. 191–196.

[16] M. Benna, N. Ariguib, A. Magnin, F. Bergaya, J. Colloid InterfacSci. 218 (1999) 442–455.

[17] D. Tessier, G. Pedro, in: Development in Sedimentology, Proc. 7thClay Conf., Elsevier, Amsterdam, 1982, pp. 165–176.

[18] D. Tessier, in: M. De Boodt, M. Hayes, A. Herbillon (Eds.), Soil Cloids and Their Association in Aggragates, Plenum, New York, 19pp. 387–415.

[19] H. Ben Rhaiem, C.H. Pons, D. Tessier, in: L.A. Schultz, HVan Olphen, F.A. Mumpton (Eds.), Proceeding of the InternatioClay Conference, Denver, The Clay Minerals Society, 1987, pp. 2297.

[20] H. Van Damme, P. Levitz, J.J. Fripiat, J.F. Alcover, L. GatineauBergaya, in: N. Boccara, M. Daoud (Eds.), Physics of Finely DividMatter, Springer-Verlag, Berlin, 1985, pp. 24–30.

,[21] L. Loeber, Etude de la structure des cakes d’argile formés surparois des puits au cours du forage, Thèse de doctorat, Université d’Oléans, 1992.

[22] Y. Li. Filtrations statiques et dynamiques de différents systèmes argielectrolytes, polymères, Thèse de l’université d’Orléans, 1996.

[23] D. Heath, T.F. Tadros, J. Colloid Interface Sci. 93 (1983) 307–319[24] H. Heller, R. Keren, Clay Miner. Soc. 49 (4) (2001) 286–291.[25] L.T. Lee, R. Rahbari, J. Lecourtier, G. Chauveteau, J. Colloid Inter

Sci. 147 (2) (1991) 351–358.[26] G. Lagaly, Appl. Clay Sci. 15 (1999) 1–9.[27] U. Cartalos, P. Baylocq, J. Lecourtier, J.M. Piau, Rev. Inst. Fr. Pé

(1997) 3.[28] Z. Zhou, P.J. Scales, D.V. Boger, Chem. Eng. Sci. 56 (2001) 29

2920.[29] Th.F. Tadros, Adv. Colloid Interface Sci. 68 (1996) 97–200.[30] R.E. Grim, Clay Mineralogy, McGraw–Hill, New York, 1953.[31] C.S. Calvert, D.A. Palkowsky, D. Pevear, in: D.R. Pevear, F.A. Mu

ton (Eds.), Quantitative Mineral Analysis of Clays, Clay Minerals Sciety, 1989, pp. 154–166.

[32] W. Xie, J. Lecourtier, Polym. Degradation Stability 38 (1992) 15164.

[33] N.J. Alderman, G.H. Meeten, J.D. Sherwood, J. Non-Newtonian FluMech. 39 (1991) 291–310.

[34] A. Magnin, J.M. Piau, J. Non-Newtonian Fluid Mech. 36 (1990) 85108.

[35] H.A. Barnes, Q.D. Nguyen, J.Non-Newtonian Fluid Mech. 98 (20011–14.

[36] M. Benna, Physico-chimie rhéologie et filtration statique de bentonites, application aux argiles de Berka et de Haidoudi, Thèsedoctorat, Université de Tunis el Manar, 2001.

[37] M.M. Ramos-Tejada, F.J. Arroyo, J.D.G. Duran, J. Colloid InterfSci. 235 (2001) 251–259.

[38] T. Permien, G. Lagaly, Clay Miner. 29 (1994) 751–760.[39] J. Lecourtier, G. Chauveteau, Muller, M. Anrhourrache, Int. J. B

Macromol. 8 (1986) 167.[40] T.A. Camesano, K.J. Wilkinson, Am. Chem. Soc. 2 (2001) 1184–1[41] S. Rossi, P.F. Luckham, S. Zhu, B.J. Briscoe, Th.F. Tadros, Rev.

Fr. Pét. 52 (2) (1997).[42] F. Csempesz, Chem. Eng. J. 80 (2000) 43–49.