Journal of Colloid and Interface Science 313 (2007) 305–314www.elsevier.com/locate/jcis

Thixotropic properties of aqueous suspensions containing cationic starchand aluminum magnesium hydrotalcite-like compound

Yan Li a,b, Wan-Guo Hou a,∗, Wei-Qun Zhu a

a Key Laboratory for Colloid and Interface Chemistry of Education Ministry, Shandong University, Jinan 250100, PR Chinab Department of Chemistry, Changzhi College, Changzhi 046011, PR China

Received 26 July 2006; accepted 24 March 2007

Available online 6 April 2007

Abstract

The rheological properties of aqueous suspensions consisting of cationic starch (CS) and positively charged aluminum magnesium hydrotalcite-like compound (HTlc) were investigated. Special emphasis was placed on the thixotropic phenomena. With the increase of mass ratio (R) of HTlcto CS, the equilibrium viscosity (ηeq) and the consistency coefficient (m) values of the suspensions increase in the range of neutral and alkaline pH(higher than 6.5) while decrease in the range of acid pH (lower than 6.5). With the increase of pH value, the ηeq and m values of the suspensions inthe R range of 0–0.08 studied increase initially and then decrease, appearing a maximum value at about pH 7.41±0.25. The CS/HTlc suspensionsdisplay viscid character and the yield point of the suspensions was not observed except the suspension with R = 0.08 in the pH range of 7.66–9.70, which showed a yield point and viscoelasticity. The CS/HTlc suspensions may display different thixotropic types: negative, complex orpositive thixotropy, depending on pH and R value. The thixotropic type of the CS/HTlc suspension may be transformed from negative (pure CSsolution), through complex (R = 0.02), into positive thixotropy (R = 0.05 and 0.08) with the increase of R in the studied R range of 0–0.08, andthe thixotropic strength of the suspensions increases initially and then decreases with pH value in the pH range studied. The mechanism of thethixotropic phenomenon is discussed.© 2007 Elsevier Inc. All rights reserved.

Keywords: Rheology; Thixotropy; Suspension; Cationic starch; Hydrotalcite-like compounds

1. Introduction

The term ‘thixotropy’ was introduced to describe the re-versible isothermal gel/sol/gel transformation induced by shear-ing and subsequent rest. Actually, it is a shear-thinning phenom-enon with time factor, now called positive thixotropy. Thereare many systems such as drilling mud, paint and coating andso on displaying positive thixotropy. On the contrary, nega-tive thixotropy, also called antithixotropy, is a rheological phe-nomenon characterized by a flow-induced increase of the vis-cosity in time. The phenomenon of negative thixotropy wasfirst described over 50 years ago and was observed for a widerange of polymer solutions [1–3] and some solid–water disper-sions [4,5]. In 1989, Hou et al. [6] found another thixotropicphenomenon, named complex thixotropy, in the suspension of

* Corresponding author.E-mail address: wghou@sdu.edu.cn (W.-G. Hou).

0021-9797/$ – see front matter © 2007 Elsevier Inc. All rights reserved.doi:10.1016/j.jcis.2007.03.071

aluminum magnesium hydrotalcite-like compounds (HTlc)/Na-montmorillonite, where positive thixotropic character and neg-ative thixotropic character, early and late, appeared for a givensuspension. From then on, a lot of researches [7–10] concernedwith the thixotropy of HTlc/clay suspensions have been doneand it was found that under different conditions, the suspen-sions can display different thixotropic properties, i.e. the typeof thixotropy of the suspensions may be influenced by mea-suring condition, pH and electrolyte etc. Up to now, all thecomplex thixotropic phenomena were observed in the suspen-sions consisted of HTlc having permanent positive charges andclay, montmorillonite or kaolinite, having permanent negativecharges. An interesting and important question needed to bemade clear is whether the complex thixotropy is a universal nat-ural phenomenon existing in suspensions. So, to find new sus-pensions displaying complex thixotropy and to investigate themethods controlling the thixotropic types of suspensions havebecome an important research focus. This paper will present

306 Y. Li et al. / Journal of Colloid and Interface Science 313 (2007) 305–314

the experimental results about the thixotropy of cationic starch(CS)/Al–Mg-HTlc suspensions.

Cationic starch (CS) is a positively charged polymer, anda large-scale commercial product that has widespread uses inmany fields [11–17] such as papermaking chemical, topicaldrug delivery systems and oilfield applications. HTlc is a posi-tively permanently charged inorganic materials, having the gen-eral formula [MII

1−xMIIIx (OH)2]x+[An−

x/n]x−·mH2O [18], where

MII and MIII are divalent and trivalent metal cations, respec-tively; An− is the charge compensating anions, or gallery anion;m is the number of moles of co-intercalated water per formulaweight of the compound; x is the number of moles of MIII performula weight of the compound. The HTlc is structurally char-acterized as containing brucite (magnesium hydroxide)-likelayers where some divalent metal cations have been substitutedby trivalent metal cations to form positively charged sheets.The metal cations occupy the centers of octahedra whose ver-texes contain hydroxide ions. These octahedra are connectedto each other by shared edges to form an infinite sheet. Thecationic charges in the layers have been termed permanent pos-itive charges, and they are compensated by the hydrated anionsbetween the stacked sheets. In recent years, interest has grownin the preparation, the characterization and the properties ofHTlc because HTlc may be widely utilized in many fields suchas catalysts and catalyst precursors, antacids, the preparationof pigments, the treatment of waste water, sunscreen agents,anionic exchangers, sorbents and rheology modifiers for bothaqueous and nonaqueous system [7,19–22].

Although the extensive studies on CS [11–17,23,24] andHTlc [6–10,18–22], respectively, have been reported, the rhe-ological behavior of CS/HTlc suspensions has not received at-tention. The aim of this study was to examine the rheological,especially thixotropic, properties of CS/HTlc suspensions. Spe-cial emphasis was placed on the influences of pH and mass ratio(R) of HTlc to CS on the thixotropy of the suspensions. The re-sults can lead us to understand the microstructure change withinCS/HTlc suspensions under shear and at rest, and it is helpfulto apprehend the thixotropic phenomenon more deeply.

2. Experimental

2.1. Materials

2.1.1. Cationic starchThe corn starch used in the experiments was supplied by

Xingmao Corn Developing Ltd. (Zhucheng, China) and of99.6% purity. The cationic substituent is quaternary ammo-nium, with chloride as counter-ion. The CS sample consistsof 72% of branchend amylopectin, and 28% of linear amylose.Elemental analysis of the CS sample was performed with an el-emental analyzer (VarioEL III, Elementar AnalysensystemeGMbH) to show that the percent contents of carbon, hydrogenand nitrogen are 40.70, 7.18 and 1.55, respectively. The de-gree of substituent (DS) is 0.2. Average molecular weight wasanalyzed using gel penetration chromatography (Agilent1100,America) to be 5.01 × 105 g/mol.

2.1.2. Cationic starch stock solutionThe dry CS sample was dissolved in water to a concentra-

tion of 20 g/l and stirred with a magnetic stirrer for 4 h, afterwhich it was heated at 90 ◦C for 30 min in order to dissolvesufficiently. The stock solution was stored in a refrigeratory at4 ◦C and used within 14 days in order to avoid substantial levelof degradation.

2.1.3. Aluminum magnesium hydrotalcite-like compoundMg–Al HTlc sol was synthesized by a co-precipitation

method [7]. The molar ratio of Mg/Al, pH and solid contentof the obtained HTlc sol are about 2:1, 9.50 and 12.86 wt%, re-spectively. The average diameter and isoelectric point (IEP) ofthe HTlc were determined by using zetasizer 3000 automaticparticle diameter and zeta potential determination apparatus(Malvern Instruments Ltd.) to be 105 nm and pH 11.2, respec-tively.

2.1.4. CS/HTlc suspensionsAppropriate amount of HTlc sol and deionized water were

added in 20 g/l CS stock solution, stirring vigorously to obtainthe CS/HTlc suspensions with different mass ratio (R) of HTlcto CS. The concentration of CS in all suspensions was fixedat 15 g/l, and the HTlc content changed from 0 to 0.12 wt%for suspension with R from 0 to 0.08. For simplicity, suspen-sions with R = 0 (pure CS solution), 0.02, 0.05 and 0.08 weresigned as S0, S2, S5 and S8, respectively. The pristine pH ofthe 15 g/l CS solution is 8.60, and that of CS/HTlc suspensionsis about 8.80. The expected pH values of the suspensions wereadjusted by 1 M HCl and NaOH solutions. After aged for about24 h, the pH of suspensions, which was employed in the exper-iments, was determined again before the experiments.

All other chemicals were analytical grade.

2.2. Methods

2.2.1. Rheological curve and viscosityThe rheological curves were measured by a controlled stress

rheometer (RS75, Haake Inc., Germany) equipped with the Z41concentric cylinder system at 25 ± 0.1 ◦C. After equilibration,test samples were sheared at a programmed rate increasing from0 to 1000 s−1 in 8 min to obtain flow curves. The flow curveswere evaluated by using the following rheological models:

1. Power law model:

(1)τ = mγ̇ n,

where τ is the shear stress (Pa), γ̇ is the shear rate (s−1),m is the consistency coefficient (Pa sn), and n is the fluidityindex (dimensionless).

2. Herschel–Bulkley model:

(2)τ = τ0 + mγ̇ n,

where τ0 is yield stress.

The viscosity (η) can be caculated from:

(3)η = τ/γ̇ .

Y. Li et al. / Journal of Colloid and Interface Science 313 (2007) 305–314 307

Fig. 1. Flow curves of CS/HTlc suspensions with various R values at pH 6.10 ± 0.30.

The equilibrium viscosity (ηeq), when the η value no longerchanged evidently with shear time, of CS/HTlc suspensions wasmeasured by the Haake RS75 rheometer equipped with the Z41concentric cylinder system at 25 ± 0.1 ◦C and at 3 s−1.

2.2.2. Oscillatory shear experimentOscillatory shear experiments were also performed on the

Haake RS75 rheometer equipped with the Z41 concentric cylin-der system under shear stress-controlled mode at 25 ± 0.1 ◦C.Stress sweeps were carried out at a constant frequency of 0.5 Hzfor HTlc/CS suspensions.

2.2.3. ThixotropyThixotropy of CS/HTlc suspensions was studied by moni-

toring the viscosity (η) change during the recovery process asa function of time (t ) after intensive shearing [8]. The increaseof η with t means positive thixotropy while the decrease of η

means negative thixotropy. Initial increase and then decrease ofη with t , or vice versa, means complex thixotropy. The CS/HTlcsuspensions were first sheared at a shear rate of 1000 s−1 for180 s, then the viscosity (ηt ) at different t was measured ata constant very low shear rate of 3 s−1 by the Haake RS75rheometer. The time interval from stopping vigorous shearingto accord data was strictly controlled to be 10 s. A parameter,θ , is used to describe the type and strength of thixotropy [5]:

(4)θ = G(ηt − η0),

where η0 is the viscosity at t = 0, G is a proportionality coef-ficient and define G = 1 for simplicity. When θ > 0 and the θ

value gradually increases with t , it means that the suspensiondisplays positive thixotropy; when θ < 0 and the θ value grad-ually decreases with t , it means that the suspension displaysnegative thixotropy; when the absolute value of θ increasesinitially and then decreases with t , the suspension would beconsidered as complex thixotropy.

A fresh sample was used for each measurement and was keptat rest for 30 min before determination.

2.2.4. TEM analysisTEM was performed by a JEM-100CX II model transmis-

sion electron microscope (TEM, Jeol, Japan). The sample wasprepared by dipping Formvar film-coated copper TEM gridsinto suspension, and the excess suspension were instantly re-moved using filter paper.

3. Results

3.1. Rheological characterization of CS/HTlc suspensions

The typical flow curves are shown in Figs. 1 and 2 for theCS/HTlc suspensions with various R values at pH 6.10 ± 0.30and 11.20 ± 0.20, respectively. Similar results were obtainedunder the other pH values. The apparent viscosity at a givenshear rate is the slope of flow curve at that shear rate. It can beseen from Figs. 1 and 2 that the slopes of flow curves decreasewith shear rate, i.e., the suspensions exhibit shear-thinning be-havior over entire range of shear rate studied. The power lawmodel and Herschel–Bulkley model were fitted to the exper-imental flow curves and it was observed that the power lawmodel generally gives a higher correlation coefficient (r2) val-ues, i.e., showing a better fit for data points under most circum-stances except for S8 at pH 7.66, 8.44 and 9.70 which showsbetter fit with Herschel–Bulkley model. The values of consis-tency index (m), fluidity index (n) and yield stress (τ0) obtainedby model fitting are listed in Table 1. Also the experimental ηeqvalues of the CS/HTlc suspensions at various R and pH val-ues are listed. As seen from Table 1, with the increase of pHvalue for a given R value, the ηeq and m values increase ini-tially and then decrease, appearing a maximum value at about

308 Y. Li et al. / Journal of Colloid and Interface Science 313 (2007) 305–314

Fig. 2. Flow curves of CS/HTlc suspensions with various R values at pH 11.20 ± 0.20.

Table 1Equilibrium viscosity (ηeq), yield stress (τ0), consistency index (m), fluidity index (n) of CS/HTlc suspensions at various R and pH values (values in parenthesesare standard deviations)

Sample R pH ηeq (mPa s) τ0 (Pa) m n r2

S0 0 5.90 305 / 0.718 (0.003) 0.610 (0.001) 0.9996.93 359 / 0.980 (0.003) 0.570 (0.001) 0.9997.65 400 / 1.303 (0.008) 0.542 (0.001) 0.9998.89 186 / 0.659 (0.006) 0.602 (0.001) 0.999

11.03 19.8 / 0.055 (0.001) 0.777 (0.001) 0.999

S2 0.02 5.40 390 / 0.693 (0.007) 0.617 (0.002) 0.9996.24 417 / 1.102 (0.005) 0.565 (0.001) 0.9997.56 559 / 1.513 (0.006) 0.528 (0.001) 0.9999.46 381 / 1.052 (0.003) 0.553 (0.001) 0.999

10.67 108 / 0.353 (0.002) 0.639 (0.001) 0.99911.40 27.9 / 0.055 (0.001) 0.783 (0.001) 0.999

S5 0.05 4.85 268 / 0.709 (0.002) 0.582 (0.001) 0.9996.38 343 / 0.763 (0.002) 0.584 (0.001) 0.9997.17 754 / 1.132 (0.003) 0.543 (0.001) 0.9997.92 572 / 1.367 (0.007) 0.526 (0.001) 0.9999.23 483 / 1.331 (0.004) 0.518 (0.001) 0.999

11.10 72.8 / 0.159 (0.001) 0.676 (0.001) 0.999

S8 0.08 5.85 207 / 0.528 (0.002) 0.612 (0.001) 0.9996.96 279 / 0.621 (0.002) 0.588 (0.001) 0.9997.66 801 0.649 (0.137) 0.553 (0.017) 0.635 (0.004) 0.9998.44 772 0.495 (0.102) 0.823 (0.010) 0.588 (0.002) 0.9999.70 515 0.292 (0.091) 0.534 (0.012) 0.622 (0.003) 0.999

11.12 91.5 / 0.528 (0.002) 0.612 (0.001) 0.999

pH 7.41 ± 0.25 for the suspensions in the R range of 0–0.08studied. In the range of neutral and alkaline pH (>6.5), the ηeq

value and m value increase with R, and the τ0 values were ob-served for the suspensions with R = 0.08 in the pH range of7.66–9.70; however, in the range of acid pH (<6.5), the ηeq

value and m value increase initially and then decrease withthe increase of R value, appearing a maximum value at aboutR = 0.02.

3.2. Viscoelasticity of CS/HTlc suspensions

Fig. 3 shows the dependence of storage moduli G′ and lossmoduli G′′ of CS/HTlc suspensions on R and τ at a constantpH of 7.70 ± 0.25 and a frequency (f ) of 0.5 Hz. It can beseen from Fig. 3 that for S0, S2 and S5, G′′ values are alwaysgreater than G′ values over the τ range studied, indicating thatthe viscidity of these suspensions are higher than their elasticity,

Y. Li et al. / Journal of Colloid and Interface Science 313 (2007) 305–314 309

Fig. 3. Stress dependence of G′ and G′′ for CS/HTlc suspensions with various R values at pH 7.70 ± 0.25.

Fig. 4. Stress dependence of G′ and G′′ for CS/HTlc suspension with R = 0.08 (S8) at different pH values.

or these suspensions may be considered as viscid fluids underthe condition studied; however, for S8 sample, the G′′ value islower than G′ in the τ range of 0.03–0.1 Pa and then higher thanG′ after the τ value higher than 0.1 Pa, which means that S8sample displays mainly elasticity in the τ range of 0.03–0.1 Paand viscidity after the τ value higher than 0.1 Pa. The depen-dence of G′ and G′′ of the S8 on pH at f = 0.5 Hz is presented

in Fig. 4. It can be seen from Fig. 4 that at pH 5.85 and 6.96,G′′ values are always larger than G′ values over the τ rangestudied; however, at pH 7.66 and 8.44, the G′ values are initiallyhigher than the G′′ values and then lower than those. These re-sults mean that the suspension of S8 displays mainly viscidityat pH 5.85 and 6.96 and while viscoelasticity at pH 7.66 and8.44 under the condition studied.

310 Y. Li et al. / Journal of Colloid and Interface Science 313 (2007) 305–314

Fig. 5. Thixotropic curves of pure CS solution (S0) at various pH values: (1) 5.90; (2) 6.93; (3) 7.65; (4) 8.89; (5) 10.01; (6) 11.03.

Fig. 6. Thixotropic curves of CS/HTlc suspension with R = 0.02 (S2) at various pH values: (1) 5.40; (2) 6.24; (3) 7.56; (4) 9.46; (5) 10.67; (6) 11.40.

3.3. Thixotropic behavior of CS/HTlc suspensions

The thixotropic curves for various suspensions (S0, S2, S5and S8) at different pH values are shown in Figs. 5–8. It can beseen from Fig. 5 that for pure CS solution (R = 0), the θ val-ues decrease initially and then reach a equilibrium value with t

at pH 5.90, 6.93, 7.65 and 8.89, respectively, meaning that thepure CS solution in the pH range of 5.90–8.89 displays nega-tive thixotropy; at pH 10.01 and 11.03, the θ values are almostindependent of t , meaning non-thixotropy.

For the suspension of S2, it can be seen from Fig. 6 that theθ values decrease initially and then increase with t at pH 5.40,6.24, 7.56, 9.46 and 10.67, respectively, showing a minimum

value at about 30 s, which means that the suspension in the pHrange of 5.40–10.67 displays complex thixotropy; at pH 11.40,the θ value increase with t , meaning positive thixotropy. Ev-idently, the thixotropy may be transformed from complexthixotropy into positive thixotropy with the increase of pH inthe range of 5.40–11.40.

Similar results were obtained for the suspensions of S5 andS8 (see Figs. 7 and 8). At low pH values (4.85 and 6.38 forS5, 5.85 for S8, respectively), the θ values of the suspensionsdecrease initially and then increase with t to show a mini-mum value, which means complex thixotropy. At high pH value(higher than 6.5), the θ values increase initially and then reacha equilibrium value with t , meaning positive thixotropy. Hence

Y. Li et al. / Journal of Colloid and Interface Science 313 (2007) 305–314 311

Fig. 7. Thixotropic curves of CS/HTlc suspension with R = 0.05 (S5) at various pH values: (1) 4.85; (2) 6.38; (3) 7.92; (4) 11.10. For the reason of clearness onlyfour lines were shown.

Fig. 8. Thixotropic curves of CS/HTlc suspension with R = 0.08 (S8) at various pH values: (1) 5.85; (2) 6.96; (3) 9.70; (4) 11.12. For the reason of clearness onlyfour lines were shown.

it can be concluded that for the suspensions of S5 and S8 thethixotropy may be transformed from complex thixotropy intopositive thixotropy with the increase of pH in the pH range stud-ied.

The above results show that the CS/HTlc suspensions maydisplay different thixotropic types, i.e. negative, complex andpositive thixotropy, respectively, depending on pH and R val-ues. To make a comparison among Figs. 5–8, it can be con-cluded that there is a tendency that the thixotropy of theCS/HTlc suspension may be transformed from negative (e.g.,for pure CS solution), through complex (e.g., for S2), into pos-itive thixotropy (e.g., for S5 and S8) with the increase of R

in the studied range of 0–0.08; also another tendency may beobserved that the CS/HTlc suspension tends to display com-plex thixotropy at a lower pH range and positive thixotropy at ahigher pH range.

In addition, it can be seen from Figs. 5–8 that for pure CSsolution (S0), S2, S5 and S8, all the |θ | values increase ini-tially and then decrease with the increase of pH value in thepH range studied, which means that the thixotropic strength ofthe suspensions increases initially and then decreases with theincrease of pH value in the pH range studied.

By the way, the change of θ with t at shear rates of 1and 6 s−1, respectively, were also studied and the similar re-

312 Y. Li et al. / Journal of Colloid and Interface Science 313 (2007) 305–314

sults were obtained, i.e., thixotropic type of suspensions maybe transformed from negative through complex into positivethixotropy with the increase of R and pH.

4. Discussion

CS is a polymer with positive charges, namely, positivelycharged polyelectrolyte. The statistics of polyelectrolyte chainconformation is governed by intra-chain coulombic repulsionbetween charged monomers, resulting in a more extended andswollen conformation as compared to neutral polymers [25].The interaction between CS molecules should involve: (1) zwit-terionic bonds [26], i.e., the positively charged quaternaryamines at exterior arms of two molecules may simultaneouslyattract one anion in the solution to form a bond through elec-trostatic attraction; (2) hydrogen bonds; and (3) van der Waalsattraction. It is well-known that more extended polymer molec-ular conformation and more strong interaction among polymermolecules would lead to a much high viscosity of polymer so-lution. The result that the ηeq and m values of S0 are greater inneutral pH condition than those in acidic and basic conditions(see Table 1) indicates that in a neutral solution the CS mole-cules are of most extended or stretched conformation and moststrong interaction among CS molecules. The decrease of theηeq and m values of pure CS solution in acidic and basic condi-tion maybe arise from the screening effect of HCl/NaOH elec-trolyte on the intra-chain coulombic repulsion between chargedmonomers within a CS molecule, the zwitterionic bonds andthe hydrogen bonds among the CS molecules.

For CS/HTlc suspension, the interaction between CS mole-cules and HTlc particles should involve: (1) bridging ef-fect [27], i.e., a HTlc particle may simultaneously adsorb twoor more CS molecules to form three-dimensional continuousnetwork structures over the whole system; (2) screening effectof HTlc particles on the zwitterionic bonds and the hydrogenbonds among the CS molecules. Similar effect has been pro-posed by Chen et al. [5] in the investigation of suspensionscontaining montmorillonite and hydrolyzed polyacrylamidemolecules. The bridging effect would raise the η value of thesuspension, and the screening effect would lower the η valueof the suspension. The adsorption of CS molecules on HTlcparticles was verified by adsorption experiments and FT-IR ob-servations (see supporting information), which maybe arisesfrom the hydrogen bonding between the ether groups or hy-droxyl groups of CS and the hydroxyl groups or –O− groups ofHTlc surface and the electrostatic binding between quaternaryammonium groups of CS and –O− groups of HTlc surface. Shi-razi et al. [28] had also proposed that the hydrogen bonding canform between the ether groups of hydroxyethylether and the hy-droxyl groups of kaolinite. Fig. 9 shows the TEM images of theHTlc and S8 at pH 9, it can be seen that the HTlc platelets in S8are covered with a cloudy layer, which maybe is an evidencethat CS chains are absorbed on the surface of HTlc particles.

As seen from Table 1, the ηeq and m values of the CS/HTlcsuspension with a given R increase initially and then de-crease, appearing a maximum value at near neutral condition(pH 7.41 ± 0.25), besides the influence of the changes of the

Fig. 9. TEM images of HTlc sol (a) and CS/HTlc suspension with R = 0.08(S8) at pH 9 (b).

ηeq and m values of the CS solution itself with pH, another rea-son may be considered as follows. The isoelectric point (pHIEP)of the HTlc is about 11.2 (see supporting information), hencein the studied pH range of 4.8–11.4, the lower the pH value is,the higher the net positive charge density of HTlc particles is,and the stronger the electrostatic repulsion between CS mole-cules and HTlc particles is. So, the low pH would weaken thebridging effect between CS molecules and HTlc particles andstrengthen the screening effect on the zwitterionic bonds andthe hydrogen bonds among the CS molecules, in turn reducingthe ηeq and m values of the suspension. In the alkaline con-dition, with the increase of pH the net positive charge densityof HTlc particles would decrease to reduce the electrostatic re-pulsion between CS and HTlc, and the surface density of –O−groups of HTlc would increase to strengthen the electrosta-tic binding between CS and –O− groups of HTlc, resultingin a more contracted CS-HTlc formation, this perhaps is thereason that the ηeq and m values of the CS/HTlc suspension de-creases with the pH increasing in the alkaline condition. It wasfound that the CS-HTlc suspension would flocculate when pHis higher than 11.2 (IEP). The experimental observation of theincrease of the ηeq value of the suspension with the increaseof R in the neutral and alkaline conditions (see Table 1) maybe contributed to the bridging effect between CS molecules

Y. Li et al. / Journal of Colloid and Interface Science 313 (2007) 305–314 313

and HTlc particles. In the acid condition (pH lower than 6.5),with the increase of R the initial increase of the η value of theCS/HTlc suspension may be contributed to the bridging effectbetween CS molecules and HTlc particles, while the subsequentdecrease of the η value may be contributed to the screeningeffect of HTlc particles on the zwitterionic bonds and the hy-drogen bonds among the CS molecules.

Because all CS molecules and HTlc particles are positivelycharged in the pH range studied, the strength of network formedthrough bridge effect among CS molecules and HTlc particlesmay be considered to be weak, which perhaps is the reasonof the experimental observations that the CS/HTlc suspensionsdisplayed viscid character and the yield point of the suspen-sions was not observed except the suspension with R = 0.08 inthe pH range of 7.66–9.70.

Regarding the mechanism of the thixotropic phenomenon,up to now it has not been thoroughly understood. Formerly,much work had been done on the positive thixotropy, the widelyknown mechanism is that structures existing among the parti-cles may be gradually destroyed by shearing thereby to decreasethe viscosity of the system, and after cessation of shearing thedestroyed structures may be recovered again, but the struc-ture recovery process has a relaxation time, so the viscosity ofthe system increases gradually with rest time. Most studies onthe negative thixotropy are concerned with polymer solutions[1,2] and polymer–solid suspensions [5], many theories suchas “aggregation theory,” “crystallization theory,” “network the-ory” and “shield effect” have been proposed [1,2,5]. Consultingthese theories, the experimental observations on the thixotropicphenomenon of the CS/HTlc suspensions in this study may beexplained as follows.

For pure CS solution (see Fig. 5), the negative thixotropymay be contributed to the formation of the shear-induced aggre-gation or supermolecular structures in the CS solution [5]. Un-der the shearing, some aggregation or supermolecular structuresamong CS molecules would be induced, leading to an increaseof the η value of the CS solution, which was verified by the ex-perimental result that the η value of the CS solution increasedwith shearing time at 1000 s−1 (see supporting information).After cessation of shearing, the shear-induced structures amongCS molecules would be gradually dissolved, making the η valueof the CS solution decrease gradually with t .

For CS/HTlc suspension, it may be supposed that adsorbedCS molecules on HTlc particles may be desorbed from HTlcparticles under the intensive shearing [5] because the adsorp-tion force of CS molecules on HTlc particles is not high, i.e.,the bridge structures between CS molecules and HTlc parti-cles can be partially destroyed by shearing, which would leadto a decrease of the η value of the suspension. Hence, for theCS/HTlc suspension, the shearing may induce two effects: Oneis the destruction of the bridge structures between CS mole-cules and HTlc particles which would lower the η value of thesuspension, and the other is the formation the shear-inducedstructures among CS molecules which would raise the η valueof the suspension. After cessation of shearing, on the one handthe gradual recovery of the destroyed bridge structures wouldmake the η value of the suspension increase gradually with t ;

on the other hand the gradual dissolve of the shear-induced su-permolecular structures among CS molecules would make the η

decrease gradually with t . The change behavior of the η valuewith t after cessation of shearing is determined by the recov-ery process of the destroyed bridge structures and the dissolveprocess of the shear-induced structures among CS molecules.When the influence strength of the recovery effect is similar tothat of the dissolve effect, it is an acceptable consideration thatthe dissolve of the shear-induced structures is a rapider processthan the recovery of the destroyed bridge structures; thus, inthe initial time stage after cessation of shearing the η value ofthe CS/HTlc suspension would decrease with t mainly arisingfrom the dissolve effect, while afterwards that would change toincrease with t mainly arising from the recovery effect, i.e., thesuspension would display complex thixotropy. When the influ-ence strength of the recovery effect is evidently higher than thatof the dissolve effect, the η value of the CS/HTlc suspensionwould gradually increase with t after cessation of shearing, i.e.,the suspension would display positive thixotropy. Contrarily,when the influence strength of the recovery effect is evidentlylower than that of the dissolve effect, the η value of the CS/HTlcsuspension would gradually decrease with t , i.e., the suspensionwould display negative thixotropy.

The higher the R value of the CS/HTlc suspension is, thestronger the influence of the recovery of the destroyed bridgestructures on the η value of the suspension after cessation ofshearing is. For the CS/HTlc suspension with low R value (forexample, S2), the influence strength of the recovery effect ofthe destroyed bridge structures on the η value of the suspensionis perhaps similar to that of the dissolve effect of the shear-induced structures after cessation of shearing, hence the S2 witha relatively low R value of 0.02 displays complex thixotropy(see Fig. 6). For the CS/HTlc suspension with high R value (forexample, S5 and S8), as seen from Table 1, in the pH range ofhigher than 6.5 the bridging effect is relatively strong, so the in-fluence strength of the recovery effect of the destroyed bridgestructures on the η value of the suspension would be higher thanthat of the dissolve effect of the shear-induced structures aftercessation of shearing, this is the reason that the S5 and S8 dis-play positive thixotropy in the pH range of higher than 6.5 (seeFigs. 7 and 8); however, as seen from Table 1, in the pH rangeof lower than 6.5 the bridging effect is relatively weak, so theinfluence strength of the recovery effect of the destroyed bridgestructures on the η value of the suspension may be similar tothat of the dissolve effect of the shear-induced structures aftercessation of shearing, which makes the S5 and S8 display com-plex thixotropy in the pH range of lower than 6.5 (see Figs. 7and 8). The increase of R value may enhance the bridging effectwithin the CS/HTlc suspension so that the type of the suspen-sions may be transformed from negative (for pure CS solution)through complex (for S2) to positive thixotropy (for S5 and S8)with the increase of R.

With the increase of pH, the strength of the microstructuresin the suspension increases initially and then decreases, just asseen from Table 1, the shear-induced change range of the η

value of the suspension would increase initially and then de-

314 Y. Li et al. / Journal of Colloid and Interface Science 313 (2007) 305–314

crease, this perhaps is the reason that the thixotropic strengthof the suspension increases initially and then decreases with theincrease of pH.

5. Conclusion

The rheological properties of aqueous suspensions consist-ing of cationic starch (CS) and positively charged aluminummagnesium hydrotalcite-like compound (HTlc) were affectedby mass ratio (R) of HTlc to CS and pH. With the increaseof R, the equilibrium viscosity (ηeq) and the consistency co-efficient (m) values of the suspensions increase in the rangeof neutral and alkaline pH (higher than 6.5) while decrease inthe range of acid pH (lower than 6.5). With the increase of pHvalue, the ηeq and m values of the suspensions in the R range of0–0.08 studied increase initially and then decrease, appearing amaximum value at about pH 7.41 ± 0.25. The CS/HTlc suspen-sions display viscid character and the yield point of the suspen-sions was not observed except the suspension with R = 0.08in the pH range of 7.66–9.70, which showed a yield point andviscoelasticity. The CS/HTlc suspensions may display differ-ent thixotropic types: negative, complex or positive thixotropy,depending on pH and R value. The thixotropic type of theCS/HTlc suspension may be transformed from negative throughcomplex into positive thixotropy with the increase of R in thestudied R range of 0–0.08, and the thixotropic strength of thesuspensions increases initially and then decreases with the in-crease of pH value in the pH range studied.

Acknowledgments

This work was supported by the National Natural ScienceFund (No. 20273041 and 20573065) and Doctor’s Special Fundof Ministry of Education (No. 20040422047).

Supporting information

The online version of this article contains additional support-ing information.

Please visit DOI: 10.1016/j.jcis.2007.03.071.

References

[1] P. Bradna, O. Quadrat, Colloid Polym. Sci. 262 (1984) 189–196.[2] O. Quadrat, Adv. Colloid Interface Sci. 24 (1985) 45–75.[3] P. Bradna, O. Quadrat, D. Dupuis, Colloid Polym. Sci. 273 (1995) 642–

647.[4] R.O. Heckroodt, W. Ryan, Trans. J. Brit. Ceram. Soc. 77 (1978) 180–183.[5] Z.Q. Chen, Y.C. Xin, X.Z. Li, C.X. Lu, Acta Chim. Sinica 47 (1989) 152–

157.[6] W.G. Hou, D.J. Sun, S.H. Han, C.G. Zhang, G.T. Wang, Colloid Polym.

Sci. 276 (1998) 274–277.[7] S.P. Li, W.G. Hou, D.J. Sun, P.Z. Guo, C.X. Jia, Langmuir 19 (2003)

3172–3177.[8] S.P. Li, W.G. Hou, C.X. Jia, J.F. Hu, D.Q. Li, Colloids Surf. A 224 (2003)

149–156.[9] S.P. Li, W.G. Hou, X.N. Dai, J. Dispersion Sci. Technol. 24 (2003) 145–

152.[10] X.N. Dai, W.G. Hou, L.F. Li, Y.N. Jiao, Acta Chim. Sinica 63 (2005) 693–

696.[11] C. Xiao, S. Gao, H. Wang, L. Zhang, J. Appl. Polym. Sci. 76 (2000) 509–

515.[12] T. Tammelin, J. Merta, L.S. Johansson, P. Stenius, Langmuir 20 (2004)

10900–10909.[13] R.L. Tseng, F.C. Wu, R.S. Juang, J. Environ. Sci. Health A 34 (1999)

1815–1828.[14] N. Levy, N. Garti, S. Magdassi, Colloids Surf. A 97 (1995) 91–99.[15] G.L. Brode, E.D. Goddard, W.C. Harris, G.A. Salensky, in: C.G. Gebelein,

T.C. Cheng, V.C. Yang (Eds.), Cosmetic and Pharmaceutical Applicationsof Polymers, Plenum Press, New York, 1991, pp. 117–128.

[16] J. Mulhbachera, P. Ispas-Szaboa, V. Lenaertsb, M.A. Mateescu, J. Con-trolled Release 76 (2001) 51–58.

[17] T. Leslie, H. Xiao, M.Z. Dong, J. Pet. Sci. Eng. 46 (2005) 225–232.[18] U. Constantino, M. Casciola, L. Massinelli, M. Nocchetti, R. Vivani, Solid

State Ionics 97 (1997) 203–212.[19] B. Sels, D.D. Vos, M. Buntinx, F. Pierard, A.K. Mesmaeker, P. Jacobs,

Nature 400 (1999) 855–857.[20] H.B. Hsueh, C.Y. Chen, Polymer 44 (2003) 1151–1161.[21] D.P. Das, J. Das, K. Parida, J. Colloid Interface Sci. 261 (2003) 213–220.[22] Y. Zou, E.R. Alírio, Energy Convers. Manage. 43 (2002) 1865–1876.[23] M. Shirazi, T.G.M. van de Ven, G. Garnier, Langmuir 19 (2003) 10829–

10834.[24] C. Montri, S. Manop, Carbohydr. Polym. 61 (2005) 288–295.[25] R.N. Roland, A. David, Phys. Rep. 380 (2003) 1–95.[26] L. Larsson, Colloids Surf. B 12 (1998) 23–34.[27] G. Schmidt, A.I. Nakatani, P.D. Butler, A. Karim, C.C. Han, Macromole-

cules 33 (2000) 7219–7222.[28] M. Shirazi, T.G.M. van de Ven, G. Garnier, Langmuir 19 (2003) 10835–

10842.