Soft Matter

Cite this: Soft Matter, 2012, 8, 6120 www.rsc.org/softmatter

View Online / Journal Homepage / Table of Contents for this issue

Dynamic Article Links <

PAPER

Ergodicity breaking and aging dynamics in LaponiteMontmorillonite mixed clay dispersions

Ravi Kumar Pujala and H. B. Bohidar*
Downloaded by Jawaharlal Nehru University on 02 July 2012 Published on 01 May 2012 on http://pubs.rsc.org | doi:10.1039/C2SM25337B

Received 15th February 2012, Accepted 4th April 2012 DOI: 10.1039/c2sm25337b Ergodicity breaking and aging dynamics in a mixture of colloidal clays, Laponite (L) and Montmorillonite (MMT), were investigated. The intensity of the light scattered off L-MMT dispersions scaled with solid volume fraction (f) as, I  fa, with a 1.49, for Laponite a 0.95, for 2 104 < f < 8 103, and it remained invariant above f 8 103. For MMT, a 1.38 and no such invariance was noticed. Ergodic to non-ergodic transitions were observed in aged samples with a characteristic ergodicity breaking time, sEB z 102 105 s. Dynamic structure factor, g1(q,t) exhibited two-mode relaxation, g1(q,t) a exp(t/s1) + (1 a) exp(t/s2)b. The stretching parameter, b was found to decrease linearly with aging time, tw whereas slow mode relaxation time increased as s2  exp (btw) and s1 remained invariant of tw. Spatial dependence of s1 and s2 showed, s1,s2  q2 in the ergodic phase, considered arising from the diffusive motion. After the ergodicity breaking time (t > sEB), the slow mode relaxation was no more diffusive and revealed s2  q1, indicating dynamic arrest of the platelets in the glass phase. The temporal evolution of the slow mode relaxation time s2 in full aging regime i.e. above sEB, followed a power law growth. The sEB was found to decrease with concentration as sEB  exp(f/f0). A sharp rise in anisotropy, Dp in the L-MMT glass system occurring precisely at the ergodicity breaking time was observed with the overall aging dependence given by Dp  t0.46 w . Heterogeneity in these samples was probed from the ColeCole plot of rheology data. It is concluded that the aging dynamics was governed by the Laponite platelets with the MMT providing a suitable matrix for appropriate interparticle interactions.

I.

Introduction

Anisotropically charged colloids dispersed in water produce rich phase diagrams at room temperature solely dependent on solute concentration, which has generated considerable interest in recent times. The colloidal sols seamlessly crossover to gel or glassy states with aging, exhibiting a hierarchy of time and length scales.17 An arrested phase comprises a free-energy landscape replete with several local minima and such a system can get trapped in one of these minima from which it cannot escape. Microscopically, the motion of a colloidal particle is thus blocked or conned by the neighboring particles, which in turn are blocked by their neighbors and so on, making the system freeze. This is clearly manifested in the aging of the system. The general understanding of such systems is that initially after its formation, the system is free to access a limited part of the phase space, and can escape from the local minima through thermal
Nanomaterials and Nanocomposite Laboratory, School of Physical Sciences, Jawaharlal Nehru University, New Delhi-110067, India. E-mail: bohi0700@mail.jnu.ac.in; Fax: +91 11 2674 1837; Tel: +91 11 2670 4637 Electronic supplementary information (ESI) available. See DOI: 10.1039/c2sm25337b

activation. However, as time elapses, the system locates deeper free-energy minima states and eventually gets trapped in one of these favorable states. Consequently, the system is unable to escape from this new found minima, making further evolutions slower. Due to this, the system is unable to reach thermodynamic equilibrium and nonergodicity ensues. Such a scenario gives rise to an enormous increase in viscosity and considerable slowing down of transport coefcients of the particles, thereby causing arrest of particle motion and absence of ow. Glasses are out of equilibrium systems with their physical properties evolving continuously. Though the glassy phase states of hard-sphere colloidal systems are well described through the classical mode-coupling theory, the same for anisotropic colloids remains far from well understood. The basic difference being below the glass transition temperature, the system does not nd itself in a free-energy minimum state and hence, evolves indenitely to reach equilibrium. Such a system exhibits aging behavior which is evident in its various macroscopically observed physical parameters. For instance, the translational and rotational diffusion and viscoelastic properties begin to reveal age dependent behavior. It has been reported that the dynamics of translational and rotational diffusion in the arrested phase of clay dispersions are qualitatively similar.8 These two degrees of
This journal is The Royal Society of Chemistry 2012

6120 | Soft Matter, 2012, 8, 61206127

View Online

freedom slow down concomitant with the growth of complex viscosity of the system and a detailed study by JabbariFarouji et al. indicated that the rotational diffusivity slowed down much faster as compared to the translational diffusivity.8 Colloidal clay dispersions have emerged as model candidates to probe multiple arrested phase states. The anisotropy associated with the geometrical shape and charge distribution in clays like Laponite and Montmorillonite attributes unique properties to their dispersions. For instance, such systems contain the presence of both attractive and repulsive terms in their interactions which allows either the system to gel or form a glassy state. The phase diagram of Laponite has been widely studied and at least two arrested phases have been identied pertaining to clay concentration less and more than 2% (w/v).79 These phases depict nontrivial aging dynamics which have been extensively studied by Ruzicka et al.911 A recent review10 has concluded that multiple nonergodic gels can be identied in the Laponite phase diagram where the arrested phases pertain to equilibrium gel (at low concentration) and to Wigner glass (at high concentration). Thus, a unifying picture possibly emerges from the experimental studies reported in the literature. A review by Ruzicka and Zaccarelli10 summarizes the literature data in a phase diagram comprising of four distinguishable phase states solely dependent on the Laponite concentration, c in salt-free condition: (i) c < 1% (w/w) yields a dispersion that slowly phase separates into a clayrich and clay-poor state, (ii) 1 < c < 2% (w/w) produces an equilibrium gel that owes its origin to attractive electrostatic interactions between clay platelets, (iii) c > 2% (w/w) generates a glassy phase where repulsive interactions between platelets abound, normally called Wigner glass and (iv) for c > 3% (w/w) a nematic arrested phase exhibiting birefringence is observed. There was considerable confusion in the literature regarding the identity of the arrested phase which has been resolved conclusively by Ruzicka et al. through dilution experiments.11 They concluded that Wigner glass of Laponite would dissolve in its solvent unlike its gel phase. Though considerable effort has been made in the past to offer a systematic and comprehensive explanation to the observed aging of arrested phase of Laponite dispersions, very few studies have been undertaken to probe the same in mixed clay systems which constitutes the principal objective of this work. We have used aqueous dispersions of the popular clays Laponite (L) (aspect ratio z30) and Montmorillonite (MMT) (aspect ratio z200) mixed in a 1 : 1 weight ratio, and systematically studied their aging dynamics in the clay concentration range, 1.5 < c < 3% (w/v). Both these clays are well known in the literature as far as their physical attributes and dispersion phase characteristics are concerned. This raises the pertinent question: How does the arrested phase of the mixed system evolve and age? We address these issues herein. In our earlier studies, the phase behaviour of L-MMT clays during their pre and post gelation scenario was discussed.12,13 A characteristic new dehydration transition, in 1 : 1 L-MMT cogel, was observed at Tc z 60  C, a temperature at which the storage modulus and depolarization ratio showed sharp increase. During this process the isotropic phase turned into an anisotropic arrested phase allowing the dehydration dynamics to be universally described in the Landau theory of phase transition formalism.13 Thus, the anisotropy associated with platelet charge and structure gives rise to peculiar
This journal is The Royal Society of Chemistry 2012

dynamical properties, which are exploited in processes like drilling and in customizing products such as cement, paper, paint and composite materials.1416

II. Materials
Laponite RD and Cloisite-Na (MMT) were purchased from Southern Clay Products, USA. Laponite powder was used as procured, and the fractionated MMT was used in the experiment. Laponite RD is a fully synthetic microcrystalline clay with a cation exchange capacity of 0.75 mmol g1, and a chemical formula of Na+0.7 [(Si8 Mg5.5 Li0.3) O20(OH)4]0.7. The specic gravity of Laponite is 2.53 g cc1. The Sodium Montmorillonite (Na-MMT) clay has cation exchange capacity of 0.92 mmol g1, and the chemical formula is (Na,Ca)0.33(Al,Mg)2Si4O10(OH)2$nH2O. Na-MMT has specic gravity 2.86 g cc1. The fractionation of MMT was done as described by Raghavan et al.17 Protocol for sample preparation of the mixed dispersions was the same as that described in ref. 12. The desired samples of individual dispersions were prepared by dispersing the clay powder in deionized water at three different concentrations 1.5, 2.0 and 2.5% (w/v) followed by vigorous stirring for 2h and then the aliquots were ltered through a 0.45 mm Millipore lter. Several sets of L-MMT mixed dispersions were prepared by mixing the individual suspensions in 1 : 1 volumetric ratio and stirring the same for 10 min to generate optically clear and homogeneous dispersions. The stopping point of stirring was considered tw 0. All the samples were prepared at room temperature T 25  C where relative humidity was <40%. Samples were stored in air tight borosilicate glass vials when not in use. It was observed that in the literature both volume fraction and % (w/v) units have been used to designate clay concentration. We shall be using volume fraction to represent solid concentration mostly, but at places % (w/v) unit will be used to enable comparison with literature data.

Downloaded by Jawaharlal Nehru University on 02 July 2012 Published on 01 May 2012 on http://pubs.rsc.org | doi:10.1039/C2SM25337B

III.
(a)

Methods
Dynamic and depolarized light scattering

Dynamic light scattering (DLS) experiments were performed on a 256 channel digital correlator, (PhotoCor Instruments, USA) that was operated in the multi-tau mode (logarithmically spaced channels). The time scale spanned 8-decades, i.e. from 0.5ms to 10s. This instrument used a 35 mW linearly polarized He:Ne laser. The probe length scale is dened by the inverse of the modulus of the scattering wave vector q where the wave vector q (4pn/l)sin(q/2), the medium refractive index is n, the excitation wavelength is l (632.8 nm) and q is the scattering angle. The scattered intensity values were measured at the scattering angle 90 Further details on dynamic light scattering can be obtained from ref. 18. The instrument was used in the polarized mode to determine the particle size of clay particles. In all the experiments, the difference between the measured and calculated base line was not allowed to go beyond 0.1% and the signal modulation was maintained above 85%. The data that showed excessive baseline difference were rejected. This ensured smooth correlation curves with a well dened baseline. The measured intensity auto-correlation functions were analyzed by the CONTIN regression software to check the relaxation time
Soft Matter, 2012, 8, 61206127 | 6121

View Online

distribution function. Robustness of the tting results was decided based on two criteria: sample to sample accuracy, and data reproducibility within the same sample. Depolarized light scattering (DPLS) experiments were performed using the same scattering geometry, but with a Glan Thomson analyzer used in front of the photomultiplier tube to enable collection of polarized and depolarized components of scattered light, IVV and IVH respectively, see ref. 13 for further details. (b) Rheology Rheological measurements, using small amplitude oscillatory shear, were performed on the samples using a controlled stress AR 500 rheometer (TA Instruments, Surrey, England). For all of the tests, the real and imaginary part of complex viscosity were computed from raw dynamic oscillatory data obtained in the frequency mode of operation using TA Instrument Rheology Advantage Data Analysis software (version 3.0.1). Samples were studied by using parallel-plate geometry of radius 20mm with a truncation gap of 500 micron. The truncation gap was deliberately chosen to be more than the length scales existing in these systems. Silicon oil was used as solvent trap to prevent loss of solvent due to evaporation. Further details are provided in ref. 12 and 13.

and remains constant for f > 0.0079. But, for MMT samples, a 1.38 and no saturation was noticed. The cut off observed in both Laponite and L-MMT dispersions was the same, fcutoff 0.0079. A weak dependence on f was detected above this cut off, which may be attributed to the random arrangement of platelets in the system. These measurements were done at the initial time tw 0. The theory of scattering predicts that a dispersion containing a distribution of clusters given by cluster size distribution ni would follow19 X \ I q; f . \ I0 . ni i2 (1)
i

Downloaded by Jawaharlal Nehru University on 02 July 2012 Published on 01 May 2012 on http://pubs.rsc.org | doi:10.1039/C2SM25337B

IV. Results and discussion

(a) Concentration dependence At the outset it was felt imperative to probe the effect of concentration f on the spatial arrangement of platelets in various dispersions in the Guinier regime, qRg  1, Rg being the radius of gyration of the scattering moiety. This was achieved though the measurement of intensity of light I(q,f) scattered off these samples in the q / 0 region and the data is presented in Fig. 1. These results reveal two regimes for the Laponite and mixed dispersions, but a single regime for the MMT samples. We observed that the variation in the intensity of the light scattered off Laponite and mixed dispersions with volume fraction up to f 0.0079 followed a power law, I(q,f)  fa, where a 0.95 and 1.49 for Laponite and mixed system respectively,

Where i is the ratio of cluster to monomer molecular weight and denes the aggregation number of the cluster and <I0> is the reference intensity. Thus, the measured scattered intensity is insensitive to the geometrical structure of scattering centers in the qRg 1 regime, but it is very sensitive to their physical size. Thus, the growth of intensity can be attributed to either the association of platelets or excluded volume interactions. Milton19 through various model calculations applicable for hard sphere dispersions has shown that <I(q,f)> varies with solute concentration as: (i) when full accounting of excluded volume interactions is considered, the intensity grows linearly in the low concentration region (c < 2% (w/v)) followed by a slow variation domain when c > 5% (w/v) and (ii) when no such correction is accounted for, the said dependence is almost linear in the entire concentration region. Thus, it can be argued that the platelets of Laponite, MMT and the mixed system are associating in an excluded volume environment. Baravian et al. have found that the shear thinning of ow in dilute and semidilute clay dispersions with f < 0.0025 in low ionic strength conditions (<5 mM) can be explained on the basis of excluded volume effects.20,21 It is not possible to infer anything more from this data at this stage. A detailed study devoted to aging dynamics has been the main focus of this work. (b) Ergodicity breaking time

Fig. 1 The variation of intensity of light scattered from the dispersions of Laponite, LaponiteMMT and MMT measured at room temperature. Note the saturation at higher f for Laponite and mixed systems. The power-law exponents (a) are indicated in the gure for each curve. Solid lines are least-squares tting of data to power-law: I(q,t)  fa.

The aging behaviour and slow dynamics is best probed by dynamic light scattering experiments. The intensity auto-correlation function g2 (q,t) cannot be correlated directly to the corresponding dynamic structure factor g1 (q,t) via the Siegert relation g2 (q.t) A + B |g1 (q,t)|2, where A denes the baseline of the correlation function as, |g2(t)|t / N A, and B is the spatial coherence factor, as the system turns non-ergodic in the arrested phase. The complications are caused by the fact that the scattering centers in the arrested phase are localized near xed average positions and are able only to execute limited Brownian motion about their mean positions. A completely ergodic system is one where the time and the ensemble averages are identical, and the system is stationary implying that the process is independent of the origin of time which is adequately satised only in case of scattering from homogeneous dilute solutions. The nonergodicity problem in DLS measurements has been dealt with in several ways which include rotating the sample to probe the entire phase space, expanding the incident beam to increase the scattering volume, and extracting the non-ergodic contribution from the measured data as a heterodyne contribution. Herein, we address the problem in the following way. The normalized intensity correlation function, g2(q,t), obtained from
This journal is The Royal Society of Chemistry 2012

6122 | Soft Matter, 2012, 8, 61206127

View Online

the sample in arrested phase can be related to the dynamic structure factor, g1(q,t) as22 g2(q,t) 1 + b[2X(1 X)g1(q,t) + X2|g1(q,t)|2] (2)

Downloaded by Jawaharlal Nehru University on 02 July 2012 Published on 01 May 2012 on http://pubs.rsc.org | doi:10.1039/C2SM25337B

where b0 is the coherence factor having a maximum value of 1. In a real experiment it denes the signal modulation which is a measure of the signal-to-noise ratio in the data. The parameter X (0 # X # 1) denes the ergodicity via the amount of heterodyne contribution present in the correlation g2(q,t) data. When the value of X 1, the system is completely ergodic i.e., in the sol state and the Siegert relation is established; however in the arrested state X < 1 and the term 2X(1 X) makes a nite contribution to g2(q,t) and hence it must be accounted for. The intercept of the plot of [g2(q,t) 1] vs. delay time t at t / 0 gives b0 [2X X2] from which the value of X can be calculated if b0 is known, which is an instrumental factor. The measured intensity auto-correlation data was analyzed exactly following the description given elsewhere.22 The pre-factor of the linear term in g1(q,t) in eqn (2) is much larger than the quadratic second term. Thus g1(q,t) z [g2(q,t) 1]/[2b0 (X(1 X))] (3)

The data shown in Fig. 2 could not be tted either to a single exponential or to a single stretched exponential function with acceptable statistical accuracy. The best least-squares t was obtained when the data was tted to two different relaxation processes: a fast and a slow one. Therefore the tting expression should contain two contributions. A practical approach to describe the shape of the experimental autocorrelation function is the sum of exponential functions given by eqn (4). Therefore dynamic structure factor g1 (q,t) is described as  b   t t 1 aexp (4) g1 q; t a exp s1 s2 where a and (1 a) are the weights of the two contributions s1 and s2, which in turn are the fast and the slow mode relaxation times respectively, and b is the stretching parameter. The fast mode relaxation time s1 is related to the inverse of the short-time diffusion coefcient Ds as s1 1/Dsq2. The stretched exponential function has been reported in the past for clay dispersions, since it has been found empirically that it provides good description of the slow relaxation processes encountered in arrested systems; this will be discussed later. Fig. 3 shows the variation of the heterodyne contribution, X, with the aging time tw, and the ergodicity breaking point is dened as the time where there is a sharp change in the value of the X parameter. As the clay concentration was increased the system developed non-ergodicity rather quickly. Interestingly, the ergodicity breaking time obtained from experimental data decreased exponentially with the solid concentration as is shown in the Fig. 4. Consequently, an empirical eqn (5) can be proposed sEB  t0 exp(f/f0) (5)

The values of the heterodyne parameter X is always less than 1 except in the homogeneous dilute solutions where ergodicity is expected. The exact evaluation of the dynamic structure factor, g1(q,t) from the measured intensity auto-correlation function g2(q,t) was achieved by putting the values of b0 and X into eqn (3) and hence the non-ergodic contribution was removed, and the dynamic structure factors obtained are plotted in the Fig. 2. Fig. 2 shows the temporal evolution of g1(q,t) for, the samples, below, at and above fcutoff 0.0079. Each plot shows similar temporal evolution dynamics. It is clearly seen that g1(q,t) does not completely relax as the samples age indicating connement of platelets in the arrested phase. Thus, as the sample ages the decay of the relaxations becomes slower, meaning that platelets have reduced mobility that give rise to less rapid intensity uctuations. As the intensity starts to become uncorrelated, the system begins to show non-ergodicity.

Where t0 2.55 108 s, and f0 8.6 104. Eqn (5) denes an observation that was not reported hitherto for any clay dispersion. (c) Relaxation dynamics

The collective dynamics operative in mixed clay dispersions above the volume fraction fcutoff was probed using dynamic light

Fig. 2 The evolution of the dynamic structure factor of the mixed system shown for sample of volume fraction: fcutoff < f 0.00988 (2.5% (w/v). The solid curves represent tting of the data to eqn (4). The arrow indicates the evolution of structure factor starting tw 0; data was taken every 500 s.

Fig. 3 Semi-log plot of the heterodyne contribution parameter X shown as function of aging time, tw. The point where there is sharp decrease in the value of X is dened as the ergodicity breaking time, sEB, see the arrow. Solid lines are guide to the eye.

This journal is The Royal Society of Chemistry 2012

Soft Matter, 2012, 8, 61206127 | 6123

View Online

Downloaded by Jawaharlal Nehru University on 02 July 2012 Published on 01 May 2012 on http://pubs.rsc.org | doi:10.1039/C2SM25337B

Fig. 4 Semi-log plot of clay concentration versus the ergodicity breaking time shown for the mixed clay dispersion. Notice that non-ergodicity is achieved earlier in concentrated samples.

scattering studies. Normally a scattering system like clay glass has a complicated free-energy landscape with many local minima. Such a system is associated with a relaxation time distribution which yields a multi-modal correlation function g1(q,t). The Laplace inverse transform of the dynamic structure smax g1 q; texpt=sdt yields the relaxafactor, g1(q,t), Gq; s
smin

tion time distribution function, G(q,s).18 Now let us focus on the behavior of relaxation times of the mixed system with different solid contents. Fig. 5 shows the distribution of relaxation times as function of waiting time for various samples.

It is clear from Fig. 5 that there exists two distinct relaxation modes in the Laponite and mixed systems, whereas only a single relaxation exists in the MMT dispersion. The arrested phase in the Laponite dispersion has been reported to contain two relaxation modes.8,23,24 Furthermore, from studying the angular dependence of the scattering function g1(q,t), it was found that the fast mode relaxation time s1 varied inversely with q2 shown in Fig. 6. Therefore, this fast mode was found to be diffusive and independent of aging time. This implied that for the short times, the particles undergo normal Brownian motion.18 For the Laponite system, invariance of the diffusive fast mode relaxation with aging time has been reported earlier.8,23,24 This study conrms the same for the L-MMT samples. As was the case for s1, the relaxation time s2 was found to scale with the q2, as shown in Fig. 6, a feature again reminiscent of classical diffusion for times less than ergodicity breaking time, although the stretched exponential relaxation of slow mode shows that there are strong hydrodynamic interactions between particles. After the ergodicity breaking time, the motion of the particles are no more diffusive and begin to scale as s2  q1. Bellour et al.23 used a multispeckle technique to account for nonergodicity in Laponite glass and found identical spatial dependence of slow mode relaxation time. Kaloun et al.25 observed similar dependence in an experiment where the dynamics of tracer diffusion was probed inside Laponite glass in the linear aging regime. Kaloun et al. and others23,24 have reported two aging regimes in colloidal Laponite glass: in the rst regime, slow mode slow mode relaxation time, s2 was observed to increase exponentially with tw, followed by a full aging region where s2 was proportional to tw. The aging behaviour in L-MMT glass was examined and

Fig. 5 Plot shows the distribution of relaxation time: (a) Laponite suspension f 0.00989 at different aging time and (b) Laponite (f 0.00989), L-MMT (f 0.00989) and MMT (f 0.00874) at the initial time tw 0. Note the relative invariance of fast mode and broadening of the slow mode for aged samples. Distribution of relaxation times and the broadening in the L-MMT suspension of f 0.00989 is shown in the inset.

Fig. 6 The characteristic times, s1 and s2 of the fast and slow relaxations for L-MMT sample, shown as function of the wave vector q, for a f 0.00988 (2.5% w/v) sample. The top plot is in the sol state before ergodicity breaking and evolves to bottom plot after ergodicity breaking due to aging.

6124 | Soft Matter, 2012, 8, 61206127

This journal is The Royal Society of Chemistry 2012

View Online

Downloaded by Jawaharlal Nehru University on 02 July 2012 Published on 01 May 2012 on http://pubs.rsc.org | doi:10.1039/C2SM25337B

Fig. 7 Slow mode relaxation time shown as function of the aging time for f 0.0079 (circle) and 0.00989 (square) for L-MMT suspensions determined from dynamic structure factor data.

the temporal evolution of slow mode relaxation time s2 is presented in Fig. 7 before the ergodicity breaking time. The data allows the relaxation time s2 to be characterized as s2  s0 exp(tw/t0) (6a)

where s0  2.8 ms, t0  0.21 ms for f 0.0079 and s0  6.47 ms, t0  0.16 ms for f 0.00989. The temporal evolution of slow mode relaxation time s2 in full aging regime i.e. above sEB, follows a power law growth given as s2  twy (6b)

degree of the system. However, the low and high concentration samples are seen to exhibit distinctive behaviour as far as dynamic arrest is concerned, a phenomenon that was repeatedly noted in several measured parameters. Finally, the stretched exponent or the width parameter was found to depend on the aging time. The exponent b was found to decrease linearly between 1 and 0, as shown in Fig. 8. The aging time tw for which b has the lowest value corresponds to the ergodic to non-ergodic transition on the laboratory time scale. It was observed that both the f z fcutoff and f > fcutoff concentration samples were associated with a b value close to 0.2 at later times and at tw 0, both started with same ergodicity, b z 0.7 0.1. JabbariFarouji et al.24 observed similar results for Laponite glass. It was felt imperative to examine this behaviour for dilute systems, f < fcutoff. The representative data is shown in Fig. 9. The dynamics for f 0.0059 was found to be quite different from the dispersions at higher concentrations. The slow mode relaxation time s2 was seen to grow much faster than exponential function and stretch exponent b was also not linearly dependent on tw, which clearly revealed that the low concentration dynamics was different, which was not fully explored in the current studies. In a study by JabbariFarouji et al.,24 it was reported that Laponite system undergoes a delay between the formation of gels and glasses for the range of concentrations 1.4 < c < 2.3%, but we did not observe such behavior for the same concentrations in our mixed system. The reproducibility of the data was veried by repeating the experiments 3 times. Our mixed system shows clear regions of gel (c < 2% w/v) and glass (c > 2% w/v) states. (d) Growth of Anisotropy with aging

where y 1.1 + 0.02 for f 0.0079 and y 0.97 + 0.02 for f 0.00989. Thus, it is interesting to note that the arrested phase of the L-MMT system ages alike the Laponite glass regardless of the presence of MMT. This feature was seen during the initial aging, tw < 104 s. For the mixed system and f > fcutoff, the short time aging prevailed in the range tw < 2 103 whereas for f z fcutoff the same was observed in the time span tw < 3 104. Thus, a concentrated system aged much faster. The slow mode decay time of the correlation function is a measure of the time a particle needs to forget its initial position, these results show that very rapidly the aging freezes-in a certain

It has been reported in the literature that Laponite clay glasses exhibit anisotropy with aging. In an interesting study, Joshi et al.

Fig. 8 Variation of stretch exponent b of the slow relaxation mode as a function of the aging time for L-MMT suspensions. Solid line are leastsquares tting of data to straight lines.

Fig. 9 Characteristic time s2 and the stretch exponent b of the slow mode relaxation as function of the aging time for the dilute L-MMT suspension, f 0.0059 (f < fcutoff, corresponds 1.5% w/v).

This journal is The Royal Society of Chemistry 2012

Soft Matter, 2012, 8, 61206127 | 6125

View Online

have shown that the anisotropy propagates downwards with time starting from the interface.26 The aging dependent growth of anisotropy in the mixed glass was examined through the measurement of depolarized component of light scattered from these samples. Depolarization ratio18 is dened as follows Dp IVH IVV (7)

Downloaded by Jawaharlal Nehru University on 02 July 2012 Published on 01 May 2012 on http://pubs.rsc.org | doi:10.1039/C2SM25337B

where IVH and IW are the depolarized and polarized components of the light scattered by a medium in the plane parallel and perpendicular to the incident linearly polarized laser beam. The depolarization ratio is a quantitative measure of spatial anisotropy present in the scattering medium. Fig. 10 depicts the growth in anisotropy with waiting time for L + MMT glass; for comparison the data from the Laponite and MMT samples are shown in the same gure. The data reveals two distinct signatures: neither Laponite nor MMT dispersions showed much anisotropy whereas the L + MMT glass displayed considerable depolarized scattering during aging. This growth could be described by a power-law expression Dp  tw0.46 (8)

Fig. 11 ColeCole plot of the complex for f 0.00989 (c 2.5% (w/v)), which is plotted for different waiting times, tw where h0 * h0 /h0 max and h00 * h00 /h00 max.

Typically, in a melt, at very low frequency, viscous behaviour is observed whereas at higher frequencies elastic properties dominate. In this formalism, h*(u) h0 (u) + ih00 (u) and the low and high frequency viscosity values are given by h0 and hN respectively. The ColeCole empirical expression is written as31 h0 hN i; 0 \ a \ 1 h* hN h 1 j uscc 1a (9)

Interestingly, the initiation of rapid growth in anisotropy exactly matches with the ergodicity breaking time discussed earlier. Laponite dispersions also show a marginal increase in the anisotropy, but it is almost constant in the experimental time, as shown in the Fig. 10. We have examined a Laponite sample over a period of 20 days, and noticed very slow growth with Dp  tw0.15. It was observed that the MMT particles present in the system tried to induce isotropy and the Laponite forced the system towards anisotropy. There are few studies in the literature on the anisotropy of Laponite system.2630 The present work is the rst study on the mixed clay system where the anisotropy of the system is conclusively shown to grow exactly at the ergodicity breaking point. (e) ColeCole plot and sample heterogeneity

The phase homogeneity in polymer solutions and melts is often deduced from the ColeCole plot where the imaginary part of the complex viscosity (h00 ) is plotted as function of the real part (h0 ).

The aforesaid expression is interpreted as arising from a superposition of several Debye relaxations.32,33 The mean relaxation time is given by scc. This plot has been used extensively to map homogeneity of soft matter systems and their composites. For a homogeneous phase, the ColeCole plot is a perfect semicircle (a 0) with a well dened relaxation time. Any deviation from this shape indicates non-homogeneous dispersion and phase segregation due to immiscibility. Such phases are associated with relaxation time distributions mentioned earlier. The ColeCole plot initially (tw 50 min) was seen to be associated with a smaller slope in the low viscosity regime and an elliptical contour indicating the presence of heterogeneity in the arrested phase (Fig. 11). As the waiting time grew, the initial slope of the plots were found to increase considerably, but largely remained invariant of tw. Such a behaviour clearly refers to enhanced heterogeneity in the sample. Thus, the mixed glass of L + MMT became anisotropic and heterogeneous simultaneously with aging.

V. Conclusion
We have systematically probed the slow dynamics in an aging LaponiteMMT glass system at room temperature. In order to conrm that the arrested phase was indeed a glass phase, dilution tests were carried out on these samples following Ruzicka et al.,11 which established the presence of glass phase in our system. Interestingly, the two key characteristic parameters dening the aging of the system, the fast and slow mode relaxation time of the dynamic structure factor, and the concentration dependence of the ergodicity breaking time, exactly followed the trend observed in Laponite glass. It was observed that the ergodicity breaking time was an exponentially decaying function of clay concentration similar to Laponite glass. Similarly, the observed cross-over in the spatial dependence of slow mode relaxation time from
This journal is The Royal Society of Chemistry 2012

Fig. 10 Depolarization ratio of the Laponite (L), Laponite + MMT (f 0.00989, L + MMT) and MMT suspensions as function of the aging time tw. Note the sharp increase in Dp at tw sEB.

6126 | Soft Matter, 2012, 8, 61206127

View Online

Downloaded by Jawaharlal Nehru University on 02 July 2012 Published on 01 May 2012 on http://pubs.rsc.org | doi:10.1039/C2SM25337B

q2 to q1 with aging was earlier seen in a Laponite system. Thus, it can be unambiguously concluded that the aging pathway observed in LaponiteMMT glass qualitatively followed the same footprints as Laponite glass. It has been reported that MMT platelets are associated with a disc diameter, d z 200 nm and zeta potential, z z 30 mV whereas the same for Laponite particles is d 20 nm and z 52 mV respectively.34 The zeta potential is linearly dependent of the net surface charge to a very good approximation35 which indicates that the Laponite charge density is higher than that of MMT. Consequently, all the interplatelet interactions will be primarily dominated and governed by LaponiteLaponite electrostatic forces in the mean eld provided by the weakly charged MMT platelets. This attributes a matrix-like platform provided by MMT on which the Laponite particles strongly interact. Such a description is consistent with the experimental observations. We had reported a dehydration transition in similar systems in earlier reports where clay concentration was maintained at 3% (w/v) and referred to it as a cogel phase.12,13 Further experiments carried out in the present work identies the said phase as glass-like. The hypothesis of Laponite platelets interacting in a mean-eld environment provided by the MMT matrix needs further exploration.

Acknowledgements
RKP acknowledges receipt of senior research fellowship from Council of Scientic and Industrial Research, Government of India.

References
1 M. Dijkstra, J. P. Hansen and P. A. Madden, Phys. Rev. Lett., 1995, 75, 2236. 2 D. Bonn, H. Tanaka, G. Wegdam, H. Kellay and J. Meunier, Europhys. Lett., 1998, 45, 52. 3 P. Coussot, Q. D. Nguyen, H. T. Huynh and D. Bonn, Phys. Rev. Lett., 2002, 88, 218301. 4 B. Ruzicka, L. Zulian and G. Ruocco, Phys. Rev. Lett., 2004, 93, 258301. 5 B. Abou and F. Gallet, Phys. Rev. Lett., 2004, 93, 160603. 6 H. Z. Cummins, J. Non-Cryst. Solids, 2007, 353, 3891. 7 B. Ruzicka, E. Zaccarelli, L. Zulian, R. Angelini, M. Sztucki, A. Moussa d, T. Narayanan and F. Sciortino, Nat. Mater., 2011, 10, 56.

8 S. Jabbari-Farouji, G. H. Wegdam and D. Bonn, J. Phys.: Condens. Matter, 2004, 16, 471. 9 B. Ruzicka, L. Zulian and G. Ruocco, Langmuir, 2006, 22, 1106. 10 B. Ruzicka and E. Zaccarelli, Soft Matter, 2011, 7, 1268. 11 B. Ruzicka, L. Zulian, E. Zaccarelli, R. Angelini, M. Sztucki, A. Moussa d and G. Ruocco, Phys. Rev. Lett., 2010, 104, 085701. 12 R. K. Pujala, N. Pawar and H. B. Bohidar, Langmuir, 2011, 27, 5193. 13 R. K. Pujala, N. Pawar and H. B. Bohidar, J. Chem. Phys., 2011, 134, 194904. 14 H. van Olphen, An Introduction to clay colloid chemistry, John-Wiley, New York, 1977. 15 Laponite Technical Bulletin, Laponite Industries Limited, LI04/90/A, 1990. 16 K. Faisandier, C. H. Pons, D. Tchoubar and F. Thomas, Clays Clay Miner., 1998, 46, 636. 17 B. H. Cipriano, T. Kashiwagi, X. Zhang and S. R. Raghavan, ACS Appl. Mater. Interfaces, 2009, 1, 30. 18 B. J. Berne and R. Pecora, Dynamic Light Scattering with Applications to Chemistry, Biology and Physics: Wiley-Interscience, New York, USA, 1976. 19 A. P. Minton, Biophys. J., 2007, 93, 1321. 20 L. J. Michot, C. Baravian, I. Bihhanic, S. Maddi, C. Moyne, J. F. L. Duval, P. Levitz and P. Davidson, Langmuir, 2009, 25, 125. 21 C. Baravian, D. Vantelon and F. Thomas, Langmuir, 2003, 19, 8109. 22 T. Coviello, E. Geissler and D. Meier, Macromolecules, 1997, 30, 2008. 23 M. Bellour, A. Knaebel, J. L. Harden, F. Lequeux and J. P. Munch, Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top., 2003, 67, 031405. 24 S. Jabbari-Farouji, G. H. Wegdam and D. Bonn, Phys. Rev. Lett., 2007, 99, 065701. 25 S. Kaloun, R. Skouri, M. Skouri, J. P. Munch and F. Schosseler, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2005, 72, 011403. 26 A. Shahin, Y. M. Joshi and S. A. Ramakrishna, Langmuir, 2011, 27, 14045. 27 B. J. Lemaire, P. Panine, J. C. P. Gabriel and P. Davidson, Europhys. Lett., 2002, 59, 55. 28 A. Mourchid, A. Delville, J. Lambard, E. LeColier and P. Levitz, Langmuir, 1995, 11, 1942. vre, 29 C. Martin, F. Pignon, A. Magnin, M. Meireles, V. LeliA P. Lindner and B. Cabane, Langmuir, 2006, 22, 4065. 30 F. Pignon, M. Abyan, C. David, A. Magnin and M. Sztucki, Langmuir, DOI: 10.1021/la201492z. 31 D. W. Davidson, Can. J. Chem., 1961, 39, 571. 32 K. S. Cole and R. H. Cole, J. Chem. Phys., 1941, 9, 341. 33 T. C. Warren, J. L. Schrag and J. D. Ferry, Biopolymers, 1973, 12, 1905. 34 J. Labanda and J. Llorens, J. Colloid Interface Sci., 2005, 289, 86. 35 H. Oshima, Adv. Colloid Interface Sci., 1995, 62, 189.

This journal is The Royal Society of Chemistry 2012

Soft Matter, 2012, 8, 61206127 | 6127